#192 by GitHub stars
BuilderIO/agent-native
A framework for building agentic apps
History
Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.
View weekly changes
community-templates/account-expert/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/external-agents/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/portability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/account-expert/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/account-expert/.agents/skills/turn-into-app/SKILL.md
Week of Oct 26, 2026 · +34 / −29 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/account-expert/.agents/skills/turn-into-skill/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-expert/.agents/skills/workspace-conventions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/external-agents/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/portability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/account-tiering/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/account-tiering/.agents/skills/turn-into-app/SKILL.md
Week of Oct 26, 2026 · +34 / −29 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/account-tiering/.agents/skills/turn-into-skill/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/account-tiering/.agents/skills/workspace-conventions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/external-agents/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/portability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/call-follow-up-drafter/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/call-follow-up-drafter/.agents/skills/turn-into-app/SKILL.md
Week of Oct 26, 2026 · +34 / −29 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/call-follow-up-drafter/.agents/skills/turn-into-skill/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/call-follow-up-drafter/.agents/skills/workspace-conventions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/external-agents/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/portability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/churn-early-warning/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/churn-early-warning/.agents/skills/turn-into-app/SKILL.md
Week of Oct 26, 2026 · +34 / −29 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/churn-early-warning/.agents/skills/turn-into-skill/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/churn-early-warning/.agents/skills/workspace-conventions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/external-agents/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/portability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/demo-clip-library/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/demo-clip-library/.agents/skills/turn-into-app/SKILL.md
Week of Oct 26, 2026 · +34 / −29 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/demo-clip-library/.agents/skills/turn-into-skill/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/demo-clip-library/.agents/skills/workspace-conventions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/external-agents/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/portability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/linkedin-icp-prospect-tracker/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/linkedin-icp-prospect-tracker/.agents/skills/turn-into-app/SKILL.md
Week of Oct 26, 2026 · +34 / −29 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/linkedin-icp-prospect-tracker/.agents/skills/turn-into-skill/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-icp-prospect-tracker/.agents/skills/workspace-conventions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/external-agents/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/portability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/linkedin-signal-watch/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/linkedin-signal-watch/.agents/skills/turn-into-app/SKILL.md
Week of Oct 26, 2026 · +34 / −29 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/linkedin-signal-watch/.agents/skills/turn-into-skill/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/linkedin-signal-watch/.agents/skills/workspace-conventions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/external-agents/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/portability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/outbound-in-your-voice/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/outbound-in-your-voice/.agents/skills/turn-into-app/SKILL.md
Week of Oct 26, 2026 · +34 / −29 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/outbound-in-your-voice/.agents/skills/turn-into-skill/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/outbound-in-your-voice/.agents/skills/workspace-conventions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/external-agents/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/portability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/win-loss-memo/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/win-loss-memo/.agents/skills/turn-into-app/SKILL.md
Week of Oct 26, 2026 · +34 / −29 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
community-templates/win-loss-memo/.agents/skills/turn-into-skill/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
community-templates/win-loss-memo/.agents/skills/workspace-conventions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/builder-code-starter/.agents/skills/authentication/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/builder-code-starter/.agents/skills/internationalization/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/builder-code-starter/.agents/skills/multi-app-workspace/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/build-an-app/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/chat/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/chat/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/default/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/agent-engines/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/app-branding/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/app-permissions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/build-an-app/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/inline-embeds/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/notifications/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/progress/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/default/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/default/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/factory/.agents/skills/review-latest-feedback/SKILL.md
Week of Oct 26, 2026 · +152 / −61 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/factory/.agents/skills/review-prs/SKILL.md
Week of Oct 26, 2026 · +267 / −93 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/headless/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/headless/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/headless/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/workspace-core/.agents/skills/a2a-protocol/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/actions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/adding-a-feature/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/adding-workspace-apps/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/agent-native-docs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/agent-native-toolkit/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/automations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/build-an-app/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/client-side-routing/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/composable-mini-apps/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/context-awareness/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/customizing-agent-native/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/delegate-to-agent/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/external-agents/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/frontend-design/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/performance/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/portability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/real-time-sync/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/recurring-jobs/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/reliable-mutations/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/secrets/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/security/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/self-modifying-code/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/shadcn-ui/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/sharing/SKILL.md
Week of Oct 26, 2026 · +34 / −19 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/workspace-core/.agents/skills/storing-data/SKILL.md
Week of Oct 26, 2026 · +19 / −1 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/workspace-core/.agents/skills/turn-into-app/SKILL.md
Week of Oct 26, 2026 · +34 / −29 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
This preview is limited. See the complete change on GitHub.
packages/core/src/templates/workspace-core/.agents/skills/turn-into-skill/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/core/src/templates/workspace-core/.agents/skills/workspace-conventions/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/creative-context/docs/skills/creative-context/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/pinpoint/.agents/skills/pinpoint/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
packages/scheduling/docs/skills/availability/SKILL.md
Week of Oct 26, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
fafa62492c79 → ff01b4009091
Verified commit ff01b4009091c2140edb34d6f790ec64e7823356
git clone --filter=blob:none --no-checkout https://github.com/BuilderIO/agent-native.git .hotskills/builderio-agent-native && git -C .hotskills/builderio-agent-native checkout --detach ff01b4009091c2140edb34d6f790ec64e7823356
# Copy each selected skill directory from .hotskills/builderio-agent-native into .claude/skills/<skill-name>/git clone --filter=blob:none --no-checkout https://github.com/BuilderIO/agent-native.git .hotskills/builderio-agent-native && git -C .hotskills/builderio-agent-native checkout --detach ff01b4009091c2140edb34d6f790ec64e7823356
# Copy each selected skill directory from .hotskills/builderio-agent-native into .agents/skills/<skill-name>/git clone --filter=blob:none --no-checkout https://github.com/BuilderIO/agent-native.git .hotskills/builderio-agent-native && git -C .hotskills/builderio-agent-native checkout --detach ff01b4009091c2140edb34d6f790ec64e7823356
# Copy each selected skill directory from .hotskills/builderio-agent-native into .gemini/skills/<skill-name>/Heat
7.0- Star growth, last 7 daysAhead of 81% of tracked repositories
- Star growth, last 24 hoursAhead of 74% of tracked repositories
- Rank movement, last 7 daysAhead of 69% of tracked repositories
- Commits, last 30 daysAhead of 100% of tracked repositories
- Accounts asking about it through MCP, last 7 daysNot available
- Hot votes this weekAhead of 0% of tracked repositories
Heat never changes the order by stars, and sponsorship never touches it. How heat works
Stars over time
Individual skills in this repository
- actions
community-templates/account-expert/.agents/skills/actions - adding-a-feature
community-templates/account-expert/.agents/skills/adding-a-feature - agent-native-docs
community-templates/account-expert/.agents/skills/agent-native-docs - agent-native-toolkit
community-templates/account-expert/.agents/skills/agent-native-toolkit - client-side-routing
community-templates/account-expert/.agents/skills/client-side-routing - context-awareness
community-templates/account-expert/.agents/skills/context-awareness - customizing-agent-native
community-templates/account-expert/.agents/skills/customizing-agent-native - delegate-to-agent
community-templates/account-expert/.agents/skills/delegate-to-agent - external-agents
community-templates/account-expert/.agents/skills/external-agents - frontend-design
community-templates/account-expert/.agents/skills/frontend-design - performance
community-templates/account-expert/.agents/skills/performance - portability
community-templates/account-expert/.agents/skills/portability - real-time-sync
community-templates/account-expert/.agents/skills/real-time-sync - reliable-mutations
community-templates/account-expert/.agents/skills/reliable-mutations - secrets
community-templates/account-expert/.agents/skills/secrets - security
community-templates/account-expert/.agents/skills/security - self-modifying-code
community-templates/account-expert/.agents/skills/self-modifying-code - shadcn-ui
community-templates/account-expert/.agents/skills/shadcn-ui - sharing
community-templates/account-expert/.agents/skills/sharing - storing-data
community-templates/account-expert/.agents/skills/storing-data - turn-into-app
community-templates/account-expert/.agents/skills/turn-into-app - turn-into-skill
community-templates/account-expert/.agents/skills/turn-into-skill - workspace-conventions
community-templates/account-expert/.agents/skills/workspace-conventions - actions
community-templates/account-tiering/.agents/skills/actions - adding-a-feature
community-templates/account-tiering/.agents/skills/adding-a-feature - agent-native-docs
community-templates/account-tiering/.agents/skills/agent-native-docs - agent-native-toolkit
community-templates/account-tiering/.agents/skills/agent-native-toolkit - client-side-routing
community-templates/account-tiering/.agents/skills/client-side-routing - context-awareness
community-templates/account-tiering/.agents/skills/context-awareness - customizing-agent-native
community-templates/account-tiering/.agents/skills/customizing-agent-native - delegate-to-agent
community-templates/account-tiering/.agents/skills/delegate-to-agent - external-agents
community-templates/account-tiering/.agents/skills/external-agents - frontend-design
community-templates/account-tiering/.agents/skills/frontend-design - performance
community-templates/account-tiering/.agents/skills/performance - portability
community-templates/account-tiering/.agents/skills/portability - real-time-sync
community-templates/account-tiering/.agents/skills/real-time-sync - reliable-mutations
community-templates/account-tiering/.agents/skills/reliable-mutations - secrets
community-templates/account-tiering/.agents/skills/secrets - security
community-templates/account-tiering/.agents/skills/security - self-modifying-code
community-templates/account-tiering/.agents/skills/self-modifying-code - shadcn-ui
community-templates/account-tiering/.agents/skills/shadcn-ui - sharing
community-templates/account-tiering/.agents/skills/sharing - storing-data
community-templates/account-tiering/.agents/skills/storing-data - turn-into-app
community-templates/account-tiering/.agents/skills/turn-into-app - turn-into-skill
community-templates/account-tiering/.agents/skills/turn-into-skill - workspace-conventions
community-templates/account-tiering/.agents/skills/workspace-conventions - actions
community-templates/call-follow-up-drafter/.agents/skills/actions - adding-a-feature
community-templates/call-follow-up-drafter/.agents/skills/adding-a-feature - agent-native-docs
community-templates/call-follow-up-drafter/.agents/skills/agent-native-docs - agent-native-toolkit
community-templates/call-follow-up-drafter/.agents/skills/agent-native-toolkit - client-side-routing
community-templates/call-follow-up-drafter/.agents/skills/client-side-routing - context-awareness
community-templates/call-follow-up-drafter/.agents/skills/context-awareness - customizing-agent-native
community-templates/call-follow-up-drafter/.agents/skills/customizing-agent-native - delegate-to-agent
community-templates/call-follow-up-drafter/.agents/skills/delegate-to-agent - external-agents
community-templates/call-follow-up-drafter/.agents/skills/external-agents - frontend-design
community-templates/call-follow-up-drafter/.agents/skills/frontend-design - performance
community-templates/call-follow-up-drafter/.agents/skills/performance - portability
community-templates/call-follow-up-drafter/.agents/skills/portability - real-time-sync
community-templates/call-follow-up-drafter/.agents/skills/real-time-sync - reliable-mutations
community-templates/call-follow-up-drafter/.agents/skills/reliable-mutations - secrets
community-templates/call-follow-up-drafter/.agents/skills/secrets - security
community-templates/call-follow-up-drafter/.agents/skills/security - self-modifying-code
community-templates/call-follow-up-drafter/.agents/skills/self-modifying-code - shadcn-ui
community-templates/call-follow-up-drafter/.agents/skills/shadcn-ui - sharing
community-templates/call-follow-up-drafter/.agents/skills/sharing - storing-data
community-templates/call-follow-up-drafter/.agents/skills/storing-data - turn-into-app
community-templates/call-follow-up-drafter/.agents/skills/turn-into-app - turn-into-skill
community-templates/call-follow-up-drafter/.agents/skills/turn-into-skill - workspace-conventions
community-templates/call-follow-up-drafter/.agents/skills/workspace-conventions - actions
community-templates/churn-early-warning/.agents/skills/actions - adding-a-feature
community-templates/churn-early-warning/.agents/skills/adding-a-feature - agent-native-docs
community-templates/churn-early-warning/.agents/skills/agent-native-docs - agent-native-toolkit
community-templates/churn-early-warning/.agents/skills/agent-native-toolkit - client-side-routing
community-templates/churn-early-warning/.agents/skills/client-side-routing - context-awareness
community-templates/churn-early-warning/.agents/skills/context-awareness - customizing-agent-native
community-templates/churn-early-warning/.agents/skills/customizing-agent-native - delegate-to-agent
community-templates/churn-early-warning/.agents/skills/delegate-to-agent - external-agents
community-templates/churn-early-warning/.agents/skills/external-agents - frontend-design
community-templates/churn-early-warning/.agents/skills/frontend-design - performance
community-templates/churn-early-warning/.agents/skills/performance - portability
community-templates/churn-early-warning/.agents/skills/portability - real-time-sync
community-templates/churn-early-warning/.agents/skills/real-time-sync - reliable-mutations
community-templates/churn-early-warning/.agents/skills/reliable-mutations - secrets
community-templates/churn-early-warning/.agents/skills/secrets - security
community-templates/churn-early-warning/.agents/skills/security - self-modifying-code
community-templates/churn-early-warning/.agents/skills/self-modifying-code - shadcn-ui
community-templates/churn-early-warning/.agents/skills/shadcn-ui - sharing
community-templates/churn-early-warning/.agents/skills/sharing - storing-data
community-templates/churn-early-warning/.agents/skills/storing-data - turn-into-app
community-templates/churn-early-warning/.agents/skills/turn-into-app - turn-into-skill
community-templates/churn-early-warning/.agents/skills/turn-into-skill - workspace-conventions
community-templates/churn-early-warning/.agents/skills/workspace-conventions - actions
community-templates/demo-clip-library/.agents/skills/actions - adding-a-feature
community-templates/demo-clip-library/.agents/skills/adding-a-feature - agent-native-docs
community-templates/demo-clip-library/.agents/skills/agent-native-docs - agent-native-toolkit
community-templates/demo-clip-library/.agents/skills/agent-native-toolkit - client-side-routing
community-templates/demo-clip-library/.agents/skills/client-side-routing - context-awareness
community-templates/demo-clip-library/.agents/skills/context-awareness - customizing-agent-native
community-templates/demo-clip-library/.agents/skills/customizing-agent-native - delegate-to-agent
community-templates/demo-clip-library/.agents/skills/delegate-to-agent - external-agents
community-templates/demo-clip-library/.agents/skills/external-agents - frontend-design
community-templates/demo-clip-library/.agents/skills/frontend-design - performance
community-templates/demo-clip-library/.agents/skills/performance - portability
community-templates/demo-clip-library/.agents/skills/portability - real-time-sync
community-templates/demo-clip-library/.agents/skills/real-time-sync - reliable-mutations
community-templates/demo-clip-library/.agents/skills/reliable-mutations - secrets
community-templates/demo-clip-library/.agents/skills/secrets - security
community-templates/demo-clip-library/.agents/skills/security - self-modifying-code
community-templates/demo-clip-library/.agents/skills/self-modifying-code - shadcn-ui
community-templates/demo-clip-library/.agents/skills/shadcn-ui - sharing
community-templates/demo-clip-library/.agents/skills/sharing - storing-data
community-templates/demo-clip-library/.agents/skills/storing-data - turn-into-app
community-templates/demo-clip-library/.agents/skills/turn-into-app - turn-into-skill
community-templates/demo-clip-library/.agents/skills/turn-into-skill - workspace-conventions
community-templates/demo-clip-library/.agents/skills/workspace-conventions - actions
community-templates/linkedin-icp-prospect-tracker/.agents/skills/actions - adding-a-feature
community-templates/linkedin-icp-prospect-tracker/.agents/skills/adding-a-feature - agent-native-docs
community-templates/linkedin-icp-prospect-tracker/.agents/skills/agent-native-docs - agent-native-toolkit
community-templates/linkedin-icp-prospect-tracker/.agents/skills/agent-native-toolkit - client-side-routing
community-templates/linkedin-icp-prospect-tracker/.agents/skills/client-side-routing - context-awareness
community-templates/linkedin-icp-prospect-tracker/.agents/skills/context-awareness - customizing-agent-native
community-templates/linkedin-icp-prospect-tracker/.agents/skills/customizing-agent-native - delegate-to-agent
community-templates/linkedin-icp-prospect-tracker/.agents/skills/delegate-to-agent - external-agents
community-templates/linkedin-icp-prospect-tracker/.agents/skills/external-agents - frontend-design
community-templates/linkedin-icp-prospect-tracker/.agents/skills/frontend-design - performance
community-templates/linkedin-icp-prospect-tracker/.agents/skills/performance - portability
community-templates/linkedin-icp-prospect-tracker/.agents/skills/portability - real-time-sync
community-templates/linkedin-icp-prospect-tracker/.agents/skills/real-time-sync - reliable-mutations
community-templates/linkedin-icp-prospect-tracker/.agents/skills/reliable-mutations - secrets
community-templates/linkedin-icp-prospect-tracker/.agents/skills/secrets - security
community-templates/linkedin-icp-prospect-tracker/.agents/skills/security - self-modifying-code
community-templates/linkedin-icp-prospect-tracker/.agents/skills/self-modifying-code - shadcn-ui
community-templates/linkedin-icp-prospect-tracker/.agents/skills/shadcn-ui - sharing
community-templates/linkedin-icp-prospect-tracker/.agents/skills/sharing - storing-data
community-templates/linkedin-icp-prospect-tracker/.agents/skills/storing-data - turn-into-app
community-templates/linkedin-icp-prospect-tracker/.agents/skills/turn-into-app - turn-into-skill
community-templates/linkedin-icp-prospect-tracker/.agents/skills/turn-into-skill - workspace-conventions
community-templates/linkedin-icp-prospect-tracker/.agents/skills/workspace-conventions - actions
community-templates/linkedin-signal-watch/.agents/skills/actions - adding-a-feature
community-templates/linkedin-signal-watch/.agents/skills/adding-a-feature - agent-native-docs
community-templates/linkedin-signal-watch/.agents/skills/agent-native-docs - agent-native-toolkit
community-templates/linkedin-signal-watch/.agents/skills/agent-native-toolkit - client-side-routing
community-templates/linkedin-signal-watch/.agents/skills/client-side-routing - context-awareness
community-templates/linkedin-signal-watch/.agents/skills/context-awareness - customizing-agent-native
community-templates/linkedin-signal-watch/.agents/skills/customizing-agent-native - delegate-to-agent
community-templates/linkedin-signal-watch/.agents/skills/delegate-to-agent - external-agents
community-templates/linkedin-signal-watch/.agents/skills/external-agents - frontend-design
community-templates/linkedin-signal-watch/.agents/skills/frontend-design - performance
community-templates/linkedin-signal-watch/.agents/skills/performance - portability
community-templates/linkedin-signal-watch/.agents/skills/portability - real-time-sync
community-templates/linkedin-signal-watch/.agents/skills/real-time-sync - reliable-mutations
community-templates/linkedin-signal-watch/.agents/skills/reliable-mutations - secrets
community-templates/linkedin-signal-watch/.agents/skills/secrets - security
community-templates/linkedin-signal-watch/.agents/skills/security - self-modifying-code
community-templates/linkedin-signal-watch/.agents/skills/self-modifying-code - shadcn-ui
community-templates/linkedin-signal-watch/.agents/skills/shadcn-ui - sharing
community-templates/linkedin-signal-watch/.agents/skills/sharing - storing-data
community-templates/linkedin-signal-watch/.agents/skills/storing-data - turn-into-app
community-templates/linkedin-signal-watch/.agents/skills/turn-into-app - turn-into-skill
community-templates/linkedin-signal-watch/.agents/skills/turn-into-skill - workspace-conventions
community-templates/linkedin-signal-watch/.agents/skills/workspace-conventions - actions
community-templates/outbound-in-your-voice/.agents/skills/actions - adding-a-feature
community-templates/outbound-in-your-voice/.agents/skills/adding-a-feature - agent-native-docs
community-templates/outbound-in-your-voice/.agents/skills/agent-native-docs - agent-native-toolkit
community-templates/outbound-in-your-voice/.agents/skills/agent-native-toolkit - client-side-routing
community-templates/outbound-in-your-voice/.agents/skills/client-side-routing - context-awareness
community-templates/outbound-in-your-voice/.agents/skills/context-awareness - customizing-agent-native
community-templates/outbound-in-your-voice/.agents/skills/customizing-agent-native - delegate-to-agent
community-templates/outbound-in-your-voice/.agents/skills/delegate-to-agent - external-agents
community-templates/outbound-in-your-voice/.agents/skills/external-agents - frontend-design
community-templates/outbound-in-your-voice/.agents/skills/frontend-design - performance
community-templates/outbound-in-your-voice/.agents/skills/performance - portability
community-templates/outbound-in-your-voice/.agents/skills/portability - real-time-sync
community-templates/outbound-in-your-voice/.agents/skills/real-time-sync - reliable-mutations
community-templates/outbound-in-your-voice/.agents/skills/reliable-mutations - secrets
community-templates/outbound-in-your-voice/.agents/skills/secrets - security
community-templates/outbound-in-your-voice/.agents/skills/security - self-modifying-code
community-templates/outbound-in-your-voice/.agents/skills/self-modifying-code - shadcn-ui
community-templates/outbound-in-your-voice/.agents/skills/shadcn-ui - sharing
community-templates/outbound-in-your-voice/.agents/skills/sharing - storing-data
community-templates/outbound-in-your-voice/.agents/skills/storing-data - turn-into-app
community-templates/outbound-in-your-voice/.agents/skills/turn-into-app - turn-into-skill
community-templates/outbound-in-your-voice/.agents/skills/turn-into-skill - workspace-conventions
community-templates/outbound-in-your-voice/.agents/skills/workspace-conventions - actions
community-templates/win-loss-memo/.agents/skills/actions - adding-a-feature
community-templates/win-loss-memo/.agents/skills/adding-a-feature - agent-native-docs
community-templates/win-loss-memo/.agents/skills/agent-native-docs - agent-native-toolkit
community-templates/win-loss-memo/.agents/skills/agent-native-toolkit - client-side-routing
community-templates/win-loss-memo/.agents/skills/client-side-routing - context-awareness
community-templates/win-loss-memo/.agents/skills/context-awareness - customizing-agent-native
community-templates/win-loss-memo/.agents/skills/customizing-agent-native - delegate-to-agent
community-templates/win-loss-memo/.agents/skills/delegate-to-agent - external-agents
community-templates/win-loss-memo/.agents/skills/external-agents - frontend-design
community-templates/win-loss-memo/.agents/skills/frontend-design - performance
community-templates/win-loss-memo/.agents/skills/performance - portability
community-templates/win-loss-memo/.agents/skills/portability - real-time-sync
community-templates/win-loss-memo/.agents/skills/real-time-sync - reliable-mutations
community-templates/win-loss-memo/.agents/skills/reliable-mutations - secrets
community-templates/win-loss-memo/.agents/skills/secrets - security
community-templates/win-loss-memo/.agents/skills/security - self-modifying-code
community-templates/win-loss-memo/.agents/skills/self-modifying-code - shadcn-ui
community-templates/win-loss-memo/.agents/skills/shadcn-ui - sharing
community-templates/win-loss-memo/.agents/skills/sharing - storing-data
community-templates/win-loss-memo/.agents/skills/storing-data - turn-into-app
community-templates/win-loss-memo/.agents/skills/turn-into-app - turn-into-skill
community-templates/win-loss-memo/.agents/skills/turn-into-skill - workspace-conventions
community-templates/win-loss-memo/.agents/skills/workspace-conventions - agent-native-docs
packages/core/docs - authentication
packages/core/src/templates/builder-code-starter/.agents/skills/authentication - internationalization
packages/core/src/templates/builder-code-starter/.agents/skills/internationalization - multi-app-workspace
packages/core/src/templates/builder-code-starter/.agents/skills/multi-app-workspace - actions
packages/core/src/templates/chat/.agents/skills/actions - adding-a-feature
packages/core/src/templates/chat/.agents/skills/adding-a-feature - agent-native-docs
packages/core/src/templates/chat/.agents/skills/agent-native-docs - agent-native-toolkit
packages/core/src/templates/chat/.agents/skills/agent-native-toolkit - build-an-app
packages/core/src/templates/chat/.agents/skills/build-an-app - client-side-routing
packages/core/src/templates/chat/.agents/skills/client-side-routing - context-awareness
packages/core/src/templates/chat/.agents/skills/context-awareness - customizing-agent-native
packages/core/src/templates/chat/.agents/skills/customizing-agent-native - delegate-to-agent
packages/core/src/templates/chat/.agents/skills/delegate-to-agent - frontend-design
packages/core/src/templates/chat/.agents/skills/frontend-design - performance
packages/core/src/templates/chat/.agents/skills/performance - real-time-sync
packages/core/src/templates/chat/.agents/skills/real-time-sync - reliable-mutations
packages/core/src/templates/chat/.agents/skills/reliable-mutations - secrets
packages/core/src/templates/chat/.agents/skills/secrets - security
packages/core/src/templates/chat/.agents/skills/security - shadcn-ui
packages/core/src/templates/chat/.agents/skills/shadcn-ui - sharing
packages/core/src/templates/chat/.agents/skills/sharing - storing-data
packages/core/src/templates/chat/.agents/skills/storing-data - actions
packages/core/src/templates/default/.agents/skills/actions - adding-a-feature
packages/core/src/templates/default/.agents/skills/adding-a-feature - agent-engines
packages/core/src/templates/default/.agents/skills/agent-engines - agent-native-docs
packages/core/src/templates/default/.agents/skills/agent-native-docs - agent-native-toolkit
packages/core/src/templates/default/.agents/skills/agent-native-toolkit - app-branding
packages/core/src/templates/default/.agents/skills/app-branding - app-permissions
packages/core/src/templates/default/.agents/skills/app-permissions - build-an-app
packages/core/src/templates/default/.agents/skills/build-an-app - client-side-routing
packages/core/src/templates/default/.agents/skills/client-side-routing - context-awareness
packages/core/src/templates/default/.agents/skills/context-awareness - customizing-agent-native
packages/core/src/templates/default/.agents/skills/customizing-agent-native - delegate-to-agent
packages/core/src/templates/default/.agents/skills/delegate-to-agent - frontend-design
packages/core/src/templates/default/.agents/skills/frontend-design - inline-embeds
packages/core/src/templates/default/.agents/skills/inline-embeds - notifications
packages/core/src/templates/default/.agents/skills/notifications - performance
packages/core/src/templates/default/.agents/skills/performance - progress
packages/core/src/templates/default/.agents/skills/progress - real-time-sync
packages/core/src/templates/default/.agents/skills/real-time-sync - reliable-mutations
packages/core/src/templates/default/.agents/skills/reliable-mutations - secrets
packages/core/src/templates/default/.agents/skills/secrets - security
packages/core/src/templates/default/.agents/skills/security - shadcn-ui
packages/core/src/templates/default/.agents/skills/shadcn-ui - sharing
packages/core/src/templates/default/.agents/skills/sharing - storing-data
packages/core/src/templates/default/.agents/skills/storing-data - review-latest-feedback
packages/core/src/templates/factory/.agents/skills/review-latest-feedback - review-prs
packages/core/src/templates/factory/.agents/skills/review-prs - actions
packages/core/src/templates/headless/.agents/skills/actions - adding-a-feature
packages/core/src/templates/headless/.agents/skills/adding-a-feature - agent-native-docs
packages/core/src/templates/headless/.agents/skills/agent-native-docs - agent-native-toolkit
packages/core/src/templates/headless/.agents/skills/agent-native-toolkit - customizing-agent-native
packages/core/src/templates/headless/.agents/skills/customizing-agent-native - delegate-to-agent
packages/core/src/templates/headless/.agents/skills/delegate-to-agent - performance
packages/core/src/templates/headless/.agents/skills/performance - reliable-mutations
packages/core/src/templates/headless/.agents/skills/reliable-mutations - secrets
packages/core/src/templates/headless/.agents/skills/secrets - security
packages/core/src/templates/headless/.agents/skills/security - self-modifying-code
packages/core/src/templates/headless/.agents/skills/self-modifying-code - sharing
packages/core/src/templates/headless/.agents/skills/sharing - storing-data
packages/core/src/templates/headless/.agents/skills/storing-data - a2a-protocol
packages/core/src/templates/workspace-core/.agents/skills/a2a-protocol - actions
packages/core/src/templates/workspace-core/.agents/skills/actions - adding-a-feature
packages/core/src/templates/workspace-core/.agents/skills/adding-a-feature - adding-workspace-apps
packages/core/src/templates/workspace-core/.agents/skills/adding-workspace-apps - agent-native-docs
packages/core/src/templates/workspace-core/.agents/skills/agent-native-docs - agent-native-toolkit
packages/core/src/templates/workspace-core/.agents/skills/agent-native-toolkit - automations
packages/core/src/templates/workspace-core/.agents/skills/automations - build-an-app
packages/core/src/templates/workspace-core/.agents/skills/build-an-app - client-side-routing
packages/core/src/templates/workspace-core/.agents/skills/client-side-routing - composable-mini-apps
packages/core/src/templates/workspace-core/.agents/skills/composable-mini-apps - context-awareness
packages/core/src/templates/workspace-core/.agents/skills/context-awareness - customizing-agent-native
packages/core/src/templates/workspace-core/.agents/skills/customizing-agent-native - delegate-to-agent
packages/core/src/templates/workspace-core/.agents/skills/delegate-to-agent - external-agents
packages/core/src/templates/workspace-core/.agents/skills/external-agents - frontend-design
packages/core/src/templates/workspace-core/.agents/skills/frontend-design - performance
packages/core/src/templates/workspace-core/.agents/skills/performance - portability
packages/core/src/templates/workspace-core/.agents/skills/portability - real-time-sync
packages/core/src/templates/workspace-core/.agents/skills/real-time-sync - recurring-jobs
packages/core/src/templates/workspace-core/.agents/skills/recurring-jobs - reliable-mutations
packages/core/src/templates/workspace-core/.agents/skills/reliable-mutations - secrets
packages/core/src/templates/workspace-core/.agents/skills/secrets - security
packages/core/src/templates/workspace-core/.agents/skills/security - self-modifying-code
packages/core/src/templates/workspace-core/.agents/skills/self-modifying-code - shadcn-ui
packages/core/src/templates/workspace-core/.agents/skills/shadcn-ui - sharing
packages/core/src/templates/workspace-core/.agents/skills/sharing - storing-data
packages/core/src/templates/workspace-core/.agents/skills/storing-data - turn-into-app
packages/core/src/templates/workspace-core/.agents/skills/turn-into-app - turn-into-skill
packages/core/src/templates/workspace-core/.agents/skills/turn-into-skill - workspace-conventions
packages/core/src/templates/workspace-core/.agents/skills/workspace-conventions - creative-context
packages/creative-context/docs/skills/creative-context - pinpoint
packages/pinpoint/.agents/skills/pinpoint - availability
packages/scheduling/docs/skills/availability - booker
packages/scheduling/docs/skills/booker - bookings
packages/scheduling/docs/skills/bookings - embeds
packages/scheduling/docs/skills/embeds - event-types
packages/scheduling/docs/skills/event-types - integrations
packages/scheduling/docs/skills/integrations - routing-forms
packages/scheduling/docs/skills/routing-forms - scheduling-basics
packages/scheduling/docs/skills/scheduling-basics - slot-engine
packages/scheduling/docs/skills/slot-engine - team-scheduling
packages/scheduling/docs/skills/team-scheduling - workflows
packages/scheduling/docs/skills/workflows - an
skills/an - assets
skills/assets - content
skills/content - context-xray
skills/context-xray - design-exploration
skills/design-exploration - turn-into-app
skills/turn-into-app - visual-edit
skills/visual-edit - visual-plan
skills/visual-plans - visual-recap
skills/visual-recap - visualize-repo
skills/visualize-repo - ab-readout
templates/analytics/.agents/skills/ab-readout - account-health
templates/analytics/.agents/skills/account-health - actions
templates/analytics/.agents/skills/actions - adding-a-feature
templates/analytics/.agents/skills/adding-a-feature - adhoc-analysis
templates/analytics/.agents/skills/adhoc-analysis - admin-surfaces
templates/analytics/.agents/skills/admin-surfaces - agent-native-docs
templates/analytics/.agents/skills/agent-native-docs - agent-native-toolkit
templates/analytics/.agents/skills/agent-native-toolkit - analysis-workspace
templates/analytics/.agents/skills/analysis-workspace - bigquery
templates/analytics/.agents/skills/bigquery - client-side-routing
templates/analytics/.agents/skills/client-side-routing - context-awareness
templates/analytics/.agents/skills/context-awareness - creative-context
templates/analytics/.agents/skills/creative-context - cross-source-analysis
templates/analytics/.agents/skills/cross-source-analysis - custom-blocks
templates/analytics/.agents/skills/custom-blocks - customizing-agent-native
templates/analytics/.agents/skills/customizing-agent-native - dashboard-management
templates/analytics/.agents/skills/dashboard-management - dashboard-ops
templates/analytics/.agents/skills/dashboard-ops - data-programs
templates/analytics/.agents/skills/data-programs - data-querying
templates/analytics/.agents/skills/data-querying - delegate-to-agent
templates/analytics/.agents/skills/delegate-to-agent - external-agents
templates/analytics/.agents/skills/external-agents - feedback-ingest
templates/analytics/.agents/skills/feedback-ingest - first-touch-radar
templates/analytics/.agents/skills/first-touch-radar - fix-brief
templates/analytics/.agents/skills/fix-brief - frontend-design
templates/analytics/.agents/skills/frontend-design - gong
templates/analytics/.agents/skills/gong - hubspot
templates/analytics/.agents/skills/hubspot - incident-investigation
templates/analytics/.agents/skills/incident-investigation - metric-watch
templates/analytics/.agents/skills/metric-watch - monitoring
templates/analytics/.agents/skills/monitoring - onboarding-journeys
templates/analytics/.agents/skills/onboarding-journeys - performance
templates/analytics/.agents/skills/performance - portability
templates/analytics/.agents/skills/portability - prometheus
templates/analytics/.agents/skills/prometheus - provider-api
templates/analytics/.agents/skills/provider-api - real-time-sync
templates/analytics/.agents/skills/real-time-sync - reliable-mutations
templates/analytics/.agents/skills/reliable-mutations - replay-triage
templates/analytics/.agents/skills/replay-triage - secrets
templates/analytics/.agents/skills/secrets - security
templates/analytics/.agents/skills/security - self-modifying-code
templates/analytics/.agents/skills/self-modifying-code - session-replay
templates/analytics/.agents/skills/session-replay - shadcn-ui
templates/analytics/.agents/skills/shadcn-ui - sharing
templates/analytics/.agents/skills/sharing - storing-data
templates/analytics/.agents/skills/storing-data - turn-into-skill
templates/analytics/.agents/skills/turn-into-skill - workspace-conventions
templates/analytics/.agents/skills/workspace-conventions - apollo
templates/analytics/.builder/skills/apollo - bigquery
templates/analytics/.builder/skills/bigquery - charts
templates/analytics/.builder/skills/charts - commonroom
templates/analytics/.builder/skills/commonroom - dataforseo
templates/analytics/.builder/skills/dataforseo - dbt
templates/analytics/.builder/skills/dbt - gcloud
templates/analytics/.builder/skills/gcloud - github
templates/analytics/.builder/skills/github - gong
templates/analytics/.builder/skills/gong - grafana
templates/analytics/.builder/skills/grafana - hubspot
templates/analytics/.builder/skills/hubspot - jira
templates/analytics/.builder/skills/jira - learn
templates/analytics/.builder/skills/learn - notion
templates/analytics/.builder/skills/notion - provider-api
templates/analytics/.builder/skills/provider-api - pylon
templates/analytics/.builder/skills/pylon - sentry
templates/analytics/.builder/skills/sentry - slack
templates/analytics/.builder/skills/slack - stripe
templates/analytics/.builder/skills/stripe - a2a-assets
templates/assets/.agents/skills/a2a-assets - actions
templates/assets/.agents/skills/actions - adding-a-feature
templates/assets/.agents/skills/adding-a-feature - agent-engines
templates/assets/.agents/skills/agent-engines - agent-native-docs
templates/assets/.agents/skills/agent-native-docs - agent-native-toolkit
templates/assets/.agents/skills/agent-native-toolkit - asset-generation
templates/assets/.agents/skills/asset-generation - assets-navigation
templates/assets/.agents/skills/assets-navigation - client-side-routing
templates/assets/.agents/skills/client-side-routing - context-awareness
templates/assets/.agents/skills/context-awareness - creative-context
templates/assets/.agents/skills/creative-context - customizing-agent-native
templates/assets/.agents/skills/customizing-agent-native - delegate-to-agent
templates/assets/.agents/skills/delegate-to-agent - external-agents
templates/assets/.agents/skills/external-agents - frontend-design
templates/assets/.agents/skills/frontend-design - image-generation
templates/assets/.agents/skills/image-generation - inline-embeds
templates/assets/.agents/skills/inline-embeds - library-management
templates/assets/.agents/skills/library-management - logo-composite
templates/assets/.agents/skills/logo-composite - notifications
templates/assets/.agents/skills/notifications - performance
templates/assets/.agents/skills/performance - portability
templates/assets/.agents/skills/portability - progress
templates/assets/.agents/skills/progress - real-time-sync
templates/assets/.agents/skills/real-time-sync - reliable-mutations
templates/assets/.agents/skills/reliable-mutations - secrets
templates/assets/.agents/skills/secrets - security
templates/assets/.agents/skills/security - self-modifying-code
templates/assets/.agents/skills/self-modifying-code - shadcn-ui
templates/assets/.agents/skills/shadcn-ui - sharing
templates/assets/.agents/skills/sharing - storing-data
templates/assets/.agents/skills/storing-data - turn-into-skill
templates/assets/.agents/skills/turn-into-skill - workspace-conventions
templates/assets/.agents/skills/workspace-conventions - actions
templates/brain/.agents/skills/actions - adding-a-feature
templates/brain/.agents/skills/adding-a-feature - agent-native-docs
templates/brain/.agents/skills/agent-native-docs - agent-native-toolkit
templates/brain/.agents/skills/agent-native-toolkit - ask-across-everything
templates/brain/.agents/skills/ask-across-everything - brain
templates/brain/.agents/skills/brain - client-side-routing
templates/brain/.agents/skills/client-side-routing - context-awareness
templates/brain/.agents/skills/context-awareness - customizing-agent-native
templates/brain/.agents/skills/customizing-agent-native - delegate-to-agent
templates/brain/.agents/skills/delegate-to-agent - external-agents
templates/brain/.agents/skills/external-agents - frontend-design
templates/brain/.agents/skills/frontend-design - ingestion-and-connectors
templates/brain/.agents/skills/ingestion-and-connectors - performance
templates/brain/.agents/skills/performance - portability
templates/brain/.agents/skills/portability - real-time-sync
templates/brain/.agents/skills/real-time-sync - reliable-mutations
templates/brain/.agents/skills/reliable-mutations - secrets
templates/brain/.agents/skills/secrets - security
templates/brain/.agents/skills/security - self-modifying-code
templates/brain/.agents/skills/self-modifying-code - shadcn-ui
templates/brain/.agents/skills/shadcn-ui - sharing
templates/brain/.agents/skills/sharing - storing-data
templates/brain/.agents/skills/storing-data - turn-into-skill
templates/brain/.agents/skills/turn-into-skill - workspace-conventions
templates/brain/.agents/skills/workspace-conventions - actions
templates/calendar/.agents/skills/actions - adding-a-feature
templates/calendar/.agents/skills/adding-a-feature - agent-native-docs
templates/calendar/.agents/skills/agent-native-docs - agent-native-toolkit
templates/calendar/.agents/skills/agent-native-toolkit - availability-booking
templates/calendar/.agents/skills/availability-booking - client-side-routing
templates/calendar/.agents/skills/client-side-routing - context-awareness
templates/calendar/.agents/skills/context-awareness - customizing-agent-native
templates/calendar/.agents/skills/customizing-agent-native - delegate-to-agent
templates/calendar/.agents/skills/delegate-to-agent - event-management
templates/calendar/.agents/skills/event-management - external-agents
templates/calendar/.agents/skills/external-agents - frontend-design
templates/calendar/.agents/skills/frontend-design - performance
templates/calendar/.agents/skills/performance - portability
templates/calendar/.agents/skills/portability - real-time-sync
templates/calendar/.agents/skills/real-time-sync - reliable-mutations
templates/calendar/.agents/skills/reliable-mutations - secrets
templates/calendar/.agents/skills/secrets - security
templates/calendar/.agents/skills/security - self-modifying-code
templates/calendar/.agents/skills/self-modifying-code - shadcn-ui
templates/calendar/.agents/skills/shadcn-ui - sharing
templates/calendar/.agents/skills/sharing - storing-data
templates/calendar/.agents/skills/storing-data - turn-into-skill
templates/calendar/.agents/skills/turn-into-skill - workspace-conventions
templates/calendar/.agents/skills/workspace-conventions - actions
templates/chat/.agents/skills/actions - adding-a-feature
templates/chat/.agents/skills/adding-a-feature - agent-native-docs
templates/chat/.agents/skills/agent-native-docs - agent-native-toolkit
templates/chat/.agents/skills/agent-native-toolkit - build-an-app
templates/chat/.agents/skills/build-an-app - client-side-routing
templates/chat/.agents/skills/client-side-routing - context-awareness
templates/chat/.agents/skills/context-awareness - customizing-agent-native
templates/chat/.agents/skills/customizing-agent-native - delegate-to-agent
templates/chat/.agents/skills/delegate-to-agent - frontend-design
templates/chat/.agents/skills/frontend-design - performance
templates/chat/.agents/skills/performance - real-time-sync
templates/chat/.agents/skills/real-time-sync - reliable-mutations
templates/chat/.agents/skills/reliable-mutations - secrets
templates/chat/.agents/skills/secrets - security
templates/chat/.agents/skills/security - shadcn-ui
templates/chat/.agents/skills/shadcn-ui - sharing
templates/chat/.agents/skills/sharing - storing-data
templates/chat/.agents/skills/storing-data - actions
templates/clips/.agents/skills/actions - agent-native-docs
templates/clips/.agents/skills/agent-native-docs - agent-native-toolkit
templates/clips/.agents/skills/agent-native-toolkit - ai-video-tools
templates/clips/.agents/skills/ai-video-tools - brain-export
templates/clips/.agents/skills/brain-export - bug-reports
templates/clips/.agents/skills/bug-reports - client-side-routing
templates/clips/.agents/skills/client-side-routing - crm-call-evidence
templates/clips/.agents/skills/crm-call-evidence - customizing-agent-native
templates/clips/.agents/skills/customizing-agent-native - delegate-to-agent
templates/clips/.agents/skills/delegate-to-agent - dictate
templates/clips/.agents/skills/dictate - external-agents
templates/clips/.agents/skills/external-agents - external-integrations
templates/clips/.agents/skills/external-integrations - frontend-design
templates/clips/.agents/skills/frontend-design - meetings
templates/clips/.agents/skills/meetings - performance
templates/clips/.agents/skills/performance - recording
templates/clips/.agents/skills/recording - reliable-mutations
templates/clips/.agents/skills/reliable-mutations - screen-memory
templates/clips/.agents/skills/screen-memory - secrets
templates/clips/.agents/skills/secrets - security
templates/clips/.agents/skills/security - self-modifying-code
templates/clips/.agents/skills/self-modifying-code - shadcn-ui
templates/clips/.agents/skills/shadcn-ui - turn-into-skill
templates/clips/.agents/skills/turn-into-skill - video-editing
templates/clips/.agents/skills/video-editing - video-sharing
templates/clips/.agents/skills/video-sharing - workspace-conventions
templates/clips/.agents/skills/workspace-conventions - actions
templates/content/.agents/skills/actions - adding-a-feature
templates/content/.agents/skills/adding-a-feature - agent-native-docs
templates/content/.agents/skills/agent-native-docs - agent-native-toolkit
templates/content/.agents/skills/agent-native-toolkit - client-side-routing
templates/content/.agents/skills/client-side-routing - content
templates/content/.agents/skills/content - content-product-development
templates/content/.agents/skills/content-product-development - context-awareness
templates/content/.agents/skills/context-awareness - creative-context
templates/content/.agents/skills/creative-context - customizing-agent-native
templates/content/.agents/skills/customizing-agent-native - delegate-to-agent
templates/content/.agents/skills/delegate-to-agent - document-editing
templates/content/.agents/skills/document-editing - external-agents
templates/content/.agents/skills/external-agents - frontend-design
templates/content/.agents/skills/frontend-design - notion-integration
templates/content/.agents/skills/notion-integration - performance
templates/content/.agents/skills/performance - portability
templates/content/.agents/skills/portability - real-time-sync
templates/content/.agents/skills/real-time-sync - reliable-mutations
templates/content/.agents/skills/reliable-mutations - secrets
templates/content/.agents/skills/secrets - security
templates/content/.agents/skills/security - self-modifying-code
templates/content/.agents/skills/self-modifying-code - shadcn-ui
templates/content/.agents/skills/shadcn-ui - sharing
templates/content/.agents/skills/sharing - storing-data
templates/content/.agents/skills/storing-data - turn-into-skill
templates/content/.agents/skills/turn-into-skill - workspace-conventions
templates/content/.agents/skills/workspace-conventions - actions
templates/crm/.agents/skills/actions - adding-a-feature
templates/crm/.agents/skills/adding-a-feature - agent-native-docs
templates/crm/.agents/skills/agent-native-docs - agent-native-toolkit
templates/crm/.agents/skills/agent-native-toolkit - client-side-routing
templates/crm/.agents/skills/client-side-routing - context-awareness
templates/crm/.agents/skills/context-awareness - crm
templates/crm/.agents/skills/crm - customizing-agent-native
templates/crm/.agents/skills/customizing-agent-native - delegate-to-agent
templates/crm/.agents/skills/delegate-to-agent - external-agents
templates/crm/.agents/skills/external-agents - frontend-design
templates/crm/.agents/skills/frontend-design - performance
templates/crm/.agents/skills/performance - portability
templates/crm/.agents/skills/portability - real-time-sync
templates/crm/.agents/skills/real-time-sync - reliable-mutations
templates/crm/.agents/skills/reliable-mutations - secrets
templates/crm/.agents/skills/secrets - security
templates/crm/.agents/skills/security - self-modifying-code
templates/crm/.agents/skills/self-modifying-code - shadcn-ui
templates/crm/.agents/skills/shadcn-ui - sharing
templates/crm/.agents/skills/sharing - storing-data
templates/crm/.agents/skills/storing-data - turn-into-skill
templates/crm/.agents/skills/turn-into-skill - workspace-conventions
templates/crm/.agents/skills/workspace-conventions - actions
templates/design/.agents/skills/actions - adding-a-feature
templates/design/.agents/skills/adding-a-feature - agent-native-docs
templates/design/.agents/skills/agent-native-docs - agent-native-toolkit
templates/design/.agents/skills/agent-native-toolkit - client-side-routing
templates/design/.agents/skills/client-side-routing - context-awareness
templates/design/.agents/skills/context-awareness - creative-context
templates/design/.agents/skills/creative-context - customizing-agent-native
templates/design/.agents/skills/customizing-agent-native - delegate-to-agent
templates/design/.agents/skills/delegate-to-agent - design-editor-architecture
templates/design/.agents/skills/design-editor-architecture - design-editor-performance
templates/design/.agents/skills/design-editor-performance - design-generation
templates/design/.agents/skills/design-generation - design-review-feedback
templates/design/.agents/skills/design-review-feedback - design-systems
templates/design/.agents/skills/design-systems - design-templates
templates/design/.agents/skills/design-templates - export-handoff
templates/design/.agents/skills/export-handoff - external-agents
templates/design/.agents/skills/external-agents - frontend-design
templates/design/.agents/skills/frontend-design - full-app-build
templates/design/.agents/skills/full-app-build - hubspot-email
templates/design/.agents/skills/hubspot-email - hubspot-email-review
templates/design/.agents/skills/hubspot-email-review - journey-storyboards
templates/design/.agents/skills/journey-storyboards - native-effects
templates/design/.agents/skills/native-effects - performance
templates/design/.agents/skills/performance - portability
templates/design/.agents/skills/portability - real-time-sync
templates/design/.agents/skills/real-time-sync - reliable-mutations
templates/design/.agents/skills/reliable-mutations - responsive-breakpoints
templates/design/.agents/skills/responsive-breakpoints - secrets
templates/design/.agents/skills/secrets - security
templates/design/.agents/skills/security - self-modifying-code
templates/design/.agents/skills/self-modifying-code - shadcn-ui
templates/design/.agents/skills/shadcn-ui - sharing
templates/design/.agents/skills/sharing - storing-data
templates/design/.agents/skills/storing-data - turn-into-skill
templates/design/.agents/skills/turn-into-skill - visual-edit
templates/design/.agents/skills/visual-edit - workspace-conventions
templates/design/.agents/skills/workspace-conventions - a2a-protocol
templates/dispatch/.agents/skills/a2a-protocol - actions
templates/dispatch/.agents/skills/actions - adding-a-feature
templates/dispatch/.agents/skills/adding-a-feature - adding-workspace-apps
templates/dispatch/.agents/skills/adding-workspace-apps - agent-native-docs
templates/dispatch/.agents/skills/agent-native-docs - agent-native-toolkit
templates/dispatch/.agents/skills/agent-native-toolkit - automations
templates/dispatch/.agents/skills/automations - build-an-app
templates/dispatch/.agents/skills/build-an-app - client-side-routing
templates/dispatch/.agents/skills/client-side-routing - composable-mini-apps
templates/dispatch/.agents/skills/composable-mini-apps - context-awareness
templates/dispatch/.agents/skills/context-awareness - customizing-agent-native
templates/dispatch/.agents/skills/customizing-agent-native - delegate-to-agent
templates/dispatch/.agents/skills/delegate-to-agent - external-agents
templates/dispatch/.agents/skills/external-agents - frontend-design
templates/dispatch/.agents/skills/frontend-design - performance
templates/dispatch/.agents/skills/performance - portability
templates/dispatch/.agents/skills/portability - real-time-sync
templates/dispatch/.agents/skills/real-time-sync - recurring-jobs
templates/dispatch/.agents/skills/recurring-jobs - reliable-mutations
templates/dispatch/.agents/skills/reliable-mutations - secrets
templates/dispatch/.agents/skills/secrets - security
templates/dispatch/.agents/skills/security - self-modifying-code
templates/dispatch/.agents/skills/self-modifying-code - shadcn-ui
templates/dispatch/.agents/skills/shadcn-ui - sharing
templates/dispatch/.agents/skills/sharing - storing-data
templates/dispatch/.agents/skills/storing-data - turn-into-app
templates/dispatch/.agents/skills/turn-into-app - turn-into-skill
templates/dispatch/.agents/skills/turn-into-skill - workspace-conventions
templates/dispatch/.agents/skills/workspace-conventions - actions
templates/factory/.agents/skills/actions - adding-a-feature
templates/factory/.agents/skills/adding-a-feature - agent-native-docs
templates/factory/.agents/skills/agent-native-docs - agent-native-toolkit
templates/factory/.agents/skills/agent-native-toolkit - client-side-routing
templates/factory/.agents/skills/client-side-routing - context-awareness
templates/factory/.agents/skills/context-awareness - customizing-agent-native
templates/factory/.agents/skills/customizing-agent-native - delegate-to-agent
templates/factory/.agents/skills/delegate-to-agent - external-agents
templates/factory/.agents/skills/external-agents - factory-workflows
templates/factory/.agents/skills/factory-workflows - frontend-design
templates/factory/.agents/skills/frontend-design - performance
templates/factory/.agents/skills/performance - portability
templates/factory/.agents/skills/portability - real-time-sync
templates/factory/.agents/skills/real-time-sync - reliable-mutations
templates/factory/.agents/skills/reliable-mutations - review-latest-feedback
templates/factory/.agents/skills/review-latest-feedback - review-prs
templates/factory/.agents/skills/review-prs - secrets
templates/factory/.agents/skills/secrets - security
templates/factory/.agents/skills/security - self-modifying-code
templates/factory/.agents/skills/self-modifying-code - shadcn-ui
templates/factory/.agents/skills/shadcn-ui - sharing
templates/factory/.agents/skills/sharing - storing-data
templates/factory/.agents/skills/storing-data - turn-into-app
templates/factory/.agents/skills/turn-into-app - turn-into-skill
templates/factory/.agents/skills/turn-into-skill - workspace-conventions
templates/factory/.agents/skills/workspace-conventions - actions
templates/forms/.agents/skills/actions - adding-a-feature
templates/forms/.agents/skills/adding-a-feature - agent-native-docs
templates/forms/.agents/skills/agent-native-docs - agent-native-toolkit
templates/forms/.agents/skills/agent-native-toolkit - client-side-routing
templates/forms/.agents/skills/client-side-routing - context-awareness
templates/forms/.agents/skills/context-awareness - customizing-agent-native
templates/forms/.agents/skills/customizing-agent-native - delegate-to-agent
templates/forms/.agents/skills/delegate-to-agent - external-agents
templates/forms/.agents/skills/external-agents - form-building
templates/forms/.agents/skills/form-building - form-publishing
templates/forms/.agents/skills/form-publishing - form-responses
templates/forms/.agents/skills/form-responses - frontend-design
templates/forms/.agents/skills/frontend-design - performance
templates/forms/.agents/skills/performance - portability
templates/forms/.agents/skills/portability - real-time-sync
templates/forms/.agents/skills/real-time-sync - reliable-mutations
templates/forms/.agents/skills/reliable-mutations - secrets
templates/forms/.agents/skills/secrets - security
templates/forms/.agents/skills/security - self-modifying-code
templates/forms/.agents/skills/self-modifying-code - shadcn-ui
templates/forms/.agents/skills/shadcn-ui - sharing
templates/forms/.agents/skills/sharing - storing-data
templates/forms/.agents/skills/storing-data - turn-into-skill
templates/forms/.agents/skills/turn-into-skill - workspace-conventions
templates/forms/.agents/skills/workspace-conventions - actions
templates/mail/.agents/skills/actions - adding-a-feature
templates/mail/.agents/skills/adding-a-feature - agent-native-docs
templates/mail/.agents/skills/agent-native-docs - agent-native-toolkit
templates/mail/.agents/skills/agent-native-toolkit - client-side-routing
templates/mail/.agents/skills/client-side-routing - contacts-and-crm
templates/mail/.agents/skills/contacts-and-crm - context-awareness
templates/mail/.agents/skills/context-awareness - customizing-agent-native
templates/mail/.agents/skills/customizing-agent-native - delegate-to-agent
templates/mail/.agents/skills/delegate-to-agent - draft-queue
templates/mail/.agents/skills/draft-queue - email-drafts
templates/mail/.agents/skills/email-drafts - external-agents
templates/mail/.agents/skills/external-agents - frontend-design
templates/mail/.agents/skills/frontend-design - inbox-automations
templates/mail/.agents/skills/inbox-automations - inbox-reads-and-triage
templates/mail/.agents/skills/inbox-reads-and-triage - mail-backends
templates/mail/.agents/skills/mail-backends - mail-interaction-parity
templates/mail/.agents/skills/mail-interaction-parity - performance
templates/mail/.agents/skills/performance - portability
templates/mail/.agents/skills/portability - provider-api-scans
templates/mail/.agents/skills/provider-api-scans - real-time-sync
templates/mail/.agents/skills/real-time-sync - reliable-mutations
templates/mail/.agents/skills/reliable-mutations - secrets
templates/mail/.agents/skills/secrets - security
templates/mail/.agents/skills/security - self-modifying-code
templates/mail/.agents/skills/self-modifying-code - shadcn-ui
templates/mail/.agents/skills/shadcn-ui - sharing
templates/mail/.agents/skills/sharing - storing-data
templates/mail/.agents/skills/storing-data - turn-into-skill
templates/mail/.agents/skills/turn-into-skill - workspace-conventions
templates/mail/.agents/skills/workspace-conventions - actions
templates/plan/.agents/skills/actions - adding-a-feature
templates/plan/.agents/skills/adding-a-feature - agent-native-docs
templates/plan/.agents/skills/agent-native-docs - agent-native-toolkit
templates/plan/.agents/skills/agent-native-toolkit - client-side-routing
templates/plan/.agents/skills/client-side-routing - context-awareness
templates/plan/.agents/skills/context-awareness - customizing-agent-native
templates/plan/.agents/skills/customizing-agent-native - delegate-to-agent
templates/plan/.agents/skills/delegate-to-agent - external-agents
templates/plan/.agents/skills/external-agents - frontend-design
templates/plan/.agents/skills/frontend-design - performance
templates/plan/.agents/skills/performance - plan-authoring-flow
templates/plan/.agents/skills/plan-authoring-flow - plan-browser-editing
templates/plan/.agents/skills/plan-browser-editing - plan-comments-and-feedback
templates/plan/.agents/skills/plan-comments-and-feedback - plan-editions
templates/plan/.agents/skills/plan-editions - plan-events
templates/plan/.agents/skills/plan-events - plan-hosted-writes
templates/plan/.agents/skills/plan-hosted-writes - plan-local-codebase-chat
templates/plan/.agents/skills/plan-local-codebase-chat - plan-review-recaps
templates/plan/.agents/skills/plan-review-recaps - plan-source-sync
templates/plan/.agents/skills/plan-source-sync - plan-version-history
templates/plan/.agents/skills/plan-version-history - portability
templates/plan/.agents/skills/portability - real-time-sync
templates/plan/.agents/skills/real-time-sync - reliable-mutations
templates/plan/.agents/skills/reliable-mutations - secrets
templates/plan/.agents/skills/secrets - security
templates/plan/.agents/skills/security - self-modifying-code
templates/plan/.agents/skills/self-modifying-code - shadcn-ui
templates/plan/.agents/skills/shadcn-ui - sharing
templates/plan/.agents/skills/sharing - storing-data
templates/plan/.agents/skills/storing-data - turn-into-skill
templates/plan/.agents/skills/turn-into-skill - visual-plan
templates/plan/.agents/skills/visual-plan - visual-recap
templates/plan/.agents/skills/visual-recap - visualize-repo
templates/plan/.agents/skills/visualize-repo - workspace-conventions
templates/plan/.agents/skills/workspace-conventions - actions
templates/slides-2/.agents/skills/actions - adding-a-feature
templates/slides-2/.agents/skills/adding-a-feature - agent-native-docs
templates/slides-2/.agents/skills/agent-native-docs - agent-native-toolkit
templates/slides-2/.agents/skills/agent-native-toolkit - analytics-data-for-decks
templates/slides-2/.agents/skills/analytics-data-for-decks - apply-slide-comments
templates/slides-2/.agents/skills/apply-slide-comments - client-side-routing
templates/slides-2/.agents/skills/client-side-routing - context-awareness
templates/slides-2/.agents/skills/context-awareness - create-deck
templates/slides-2/.agents/skills/create-deck - creative-context
templates/slides-2/.agents/skills/creative-context - customizing-agent-native
templates/slides-2/.agents/skills/customizing-agent-native - deck-management
templates/slides-2/.agents/skills/deck-management - delegate-to-agent
templates/slides-2/.agents/skills/delegate-to-agent - design-systems
templates/slides-2/.agents/skills/design-systems - external-agents
templates/slides-2/.agents/skills/external-agents - frontend-design
templates/slides-2/.agents/skills/frontend-design - image-generation-via-a2a
templates/slides-2/.agents/skills/image-generation-via-a2a - performance
templates/slides-2/.agents/skills/performance - portability
templates/slides-2/.agents/skills/portability - real-time-sync
templates/slides-2/.agents/skills/real-time-sync - reliable-mutations
templates/slides-2/.agents/skills/reliable-mutations - secrets
templates/slides-2/.agents/skills/secrets - security
templates/slides-2/.agents/skills/security - self-modifying-code
templates/slides-2/.agents/skills/self-modifying-code - shadcn-ui
templates/slides-2/.agents/skills/shadcn-ui - sharing
templates/slides-2/.agents/skills/sharing - slide-design
templates/slides-2/.agents/skills/slide-design - slide-editing
templates/slides-2/.agents/skills/slide-editing - slide-images
templates/slides-2/.agents/skills/slide-images - storing-data
templates/slides-2/.agents/skills/storing-data - turn-into-skill
templates/slides-2/.agents/skills/turn-into-skill - workspace-conventions
templates/slides-2/.agents/skills/workspace-conventions - actions
templates/slides/.agents/skills/actions - adding-a-feature
templates/slides/.agents/skills/adding-a-feature - agent-native-docs
templates/slides/.agents/skills/agent-native-docs - agent-native-toolkit
templates/slides/.agents/skills/agent-native-toolkit - analytics-data-for-decks
templates/slides/.agents/skills/analytics-data-for-decks - apply-slide-comments
templates/slides/.agents/skills/apply-slide-comments - client-side-routing
templates/slides/.agents/skills/client-side-routing - context-awareness
templates/slides/.agents/skills/context-awareness - create-deck
templates/slides/.agents/skills/create-deck - creative-context
templates/slides/.agents/skills/creative-context - customizing-agent-native
templates/slides/.agents/skills/customizing-agent-native - deck-management
templates/slides/.agents/skills/deck-management - delegate-to-agent
templates/slides/.agents/skills/delegate-to-agent - design-systems
templates/slides/.agents/skills/design-systems - external-agents
templates/slides/.agents/skills/external-agents - frontend-design
templates/slides/.agents/skills/frontend-design - image-generation-via-a2a
templates/slides/.agents/skills/image-generation-via-a2a - performance
templates/slides/.agents/skills/performance - portability
templates/slides/.agents/skills/portability - real-time-sync
templates/slides/.agents/skills/real-time-sync - reliable-mutations
templates/slides/.agents/skills/reliable-mutations - secrets
templates/slides/.agents/skills/secrets - security
templates/slides/.agents/skills/security - self-modifying-code
templates/slides/.agents/skills/self-modifying-code - shadcn-ui
templates/slides/.agents/skills/shadcn-ui - sharing
templates/slides/.agents/skills/sharing - slide-design
templates/slides/.agents/skills/slide-design - slide-editing
templates/slides/.agents/skills/slide-editing - slide-images
templates/slides/.agents/skills/slide-images - storing-data
templates/slides/.agents/skills/storing-data - turn-into-skill
templates/slides/.agents/skills/turn-into-skill - workspace-conventions
templates/slides/.agents/skills/workspace-conventions - action-reference
templates/tasks/.agents/skills/action-reference - actions
templates/tasks/.agents/skills/actions - adding-a-feature
templates/tasks/.agents/skills/adding-a-feature - agent-native-docs
templates/tasks/.agents/skills/agent-native-docs - agent-native-toolkit
templates/tasks/.agents/skills/agent-native-toolkit - agent-watchdog
templates/tasks/.agents/skills/agent-watchdog - change-summary
templates/tasks/.agents/skills/change-summary - client-side-routing
templates/tasks/.agents/skills/client-side-routing - context-awareness
templates/tasks/.agents/skills/context-awareness - custom-fields
templates/tasks/.agents/skills/custom-fields - customizing-agent-native
templates/tasks/.agents/skills/customizing-agent-native - delegate-to-agent
templates/tasks/.agents/skills/delegate-to-agent - efficient-fable
templates/tasks/.agents/skills/efficient-fable - efficient-frontier
templates/tasks/.agents/skills/efficient-frontier - external-agents
templates/tasks/.agents/skills/external-agents - frontend-design
templates/tasks/.agents/skills/frontend-design - performance
templates/tasks/.agents/skills/performance - plan-arbiter
templates/tasks/.agents/skills/plan-arbiter - plow-ahead
templates/tasks/.agents/skills/plow-ahead - portability
templates/tasks/.agents/skills/portability - quick-recap
templates/tasks/.agents/skills/quick-recap - read-the-damn-docs
templates/tasks/.agents/skills/read-the-damn-docs - real-time-sync
templates/tasks/.agents/skills/real-time-sync - reliable-mutations
templates/tasks/.agents/skills/reliable-mutations - secrets
templates/tasks/.agents/skills/secrets - security
templates/tasks/.agents/skills/security - self-modifying-code
templates/tasks/.agents/skills/self-modifying-code - shadcn-ui
templates/tasks/.agents/skills/shadcn-ui - sharing
templates/tasks/.agents/skills/sharing - stay-within-limits
templates/tasks/.agents/skills/stay-within-limits - store-conventions
templates/tasks/.agents/skills/store-conventions - storing-data
templates/tasks/.agents/skills/storing-data - task-inbox-workflow
templates/tasks/.agents/skills/task-inbox-workflow - turn-into-skill
templates/tasks/.agents/skills/turn-into-skill - visual-plan
templates/tasks/.agents/skills/visual-plan - visual-recap
templates/tasks/.agents/skills/visual-recap - workspace-conventions
templates/tasks/.agents/skills/workspace-conventions - upgrade-agent-native
templates/videos/.agents/skills/upgrade-agent-native
How it works
actions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
89,038 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
../../../../../.agents/plugins/agent-native-design/.claude-plugin/plugin.jsonasset673 bytes../../../../../.agents/plugins/agent-native-design/.mcp.jsonasset134 bytes../../../../../.agents/plugins/agent-native-design/skills/design-exploration/SKILL.mddoc6,799 bytes../../../../../.agents/plugins/agent-native-design/skills/visual-edit/README.mddoc2,249 bytes../../../../../.agents/plugins/agent-native-design/skills/visual-edit/SKILL.mddoc30,968 bytes../../../../../.agents/plugins/agent-native-turn-into-app/.claude-plugin/plugin.jsonasset864 bytes../../../../../.agents/plugins/agent-native-turn-into-app/.claude-plugin/plugin.jsonasset864 bytes../../../../../.agents/plugins/agent-native-turn-into-app/.mcp.jsonasset138 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/SKILL.mddoc16,932 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/agents/openai.yamlasset383 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/attachments.mddoc1,503 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/fresh-project.mddoc4,733 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/local-run-and-deploy.mddoc7,649 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/review-loop.mddoc19,749 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/source-brief.mddoc6,588 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/spreadsheet-source.mddoc14,401 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/ui-archetypes.mddoc19,180 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/ui-direction.mddoc19,431 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/ui-palettes.mddoc23,241 bytes../../../../../.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/where-it-runs.mddoc4,905 bytes../../../../../.agents/plugins/agent-native-visual-plans/.claude-plugin/plugin.jsonasset801 bytes../../../../../.agents/plugins/agent-native-visual-plans/.mcp.jsonasset101 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-plan/SKILL.mddoc32,771 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-plan/references/canvas.mddoc9,388 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-plan/references/connection.mddoc2,980 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-plan/references/document-quality.mddoc12,638 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-plan/references/exemplar.mddoc4,028 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-plan/references/local-files.mddoc6,509 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-plan/references/wireframe.mddoc19,471 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-recap/SKILL.mddoc33,616 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-recap/references/connection.mddoc2,980 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-recap/references/local-files.mddoc6,509 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visual-recap/references/wireframe.mddoc19,471 bytes../../../../../.agents/plugins/agent-native-visual-plans/skills/visualize-repo/SKILL.mddoc3,785 bytes../../../../../.agents/plugins/agent-native/.claude-plugin/plugin.jsonasset736 bytes../../../../../.agents/plugins/agent-native/.mcp.jsonasset138 bytes../../../../../.agents/plugins/agent-native/skills/an/SKILL.mddoc4,692 bytes../../../../../.claude-plugin/marketplace.jsonasset2,690 bytes../../../../../.claude-plugin/marketplace.jsonasset2,690 bytes../../../../../.codex/hooks.jsonasset792 bytes../../../../../.codex/hooks.jsonasset792 bytes../../../../../.codex/hooks.jsonasset792 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubexternal-agents
When your agent uses it
Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.
What it loads into context
Frontmatter
148 tokensAlways available to the agent
SKILL.md
6,857 tokensWhen the skill is activated
Reference documents
2,355 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- App-facing MCP connectors
- Provider API connectors
- How
- Do
- Don't
- Related Skills
- Blueprint installer
Files
SKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesreferences/mcp-apps-embedding.mddoc10,363 bytes
Read SKILL.md
# External Agents (MCP bridge + deep links)
## Rule
- **The connected model is the author.** Write content yourself with create/update tools; never delegate via `ask_app` or wait on an unreachable in-app form — see `DEFAULT_AGENT_NATIVE_MCP_INSTRUCTIONS`.
- **Result shape and key tools are in `actions`** (Return Values, Key Actions).
- **Artifacts fit in one call.** Up to 500,000 characters; return a `storing-data` handle.
- **One tab, one id.** WebMCP sends `X-Agent-Native-Browser-Tab`; state resolves to the driving tab (`context-awareness`).
An agent-native app is reachable by any MCP-compatible host. Keep
setup simple: for workspace or cross-app access, add one remote MCP connector:
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page
controls whether that single connector reaches all apps or only selected apps,
and Dispatch filters `list_apps`, `ask_app`, and `open_app` to the granted set.View the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubportability
When your agent uses it
How to keep template code PostgreSQL-specific and hosting-agnostic. Use when defining schemas, writing raw SQL, or creating server routes.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
956 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Database
- Hosting Agnostic
- Related Skills
Files
SKILL.mdcore4,257 bytesSKILL.mdcore4,257 bytes
Read SKILL.md
# Portability
## Rule
Templates use one PostgreSQL schema and query contract. Local PGlite and hosted
PostgreSQL share the same SQL semantics.
## Database
Use Drizzle's PostgreSQL exports directly for schemas and its query builder for reads/writes:
```ts
import { sql } from "drizzle-orm";
import {View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubturn-into-app
When your agent uses it
Turns a thread, skill, spreadsheet, or Claude/ChatGPT project into a polished, visual Agent-Native app: populated domain screens with the in-app agent working behind contextual controls. Use when a user invokes `/turn-into-app`, or asks to turn a workflow, skill, spreadsheet, or project into an app, UI, workbench, or dashboard, including from Claude or ChatGPT on the web.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
3,995 tokensWhen the skill is activated
Reference documents
33,236 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Host, source, and when to ask
- Workflow
- Final report
- Related skills
Files
SKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesagents/openai.yamlasset383 bytesreferences/attachments.mddoc1,503 bytesreferences/fresh-project.mddoc4,733 bytesreferences/local-run-and-deploy.mddoc7,649 bytesreferences/review-loop.mddoc19,749 bytesreferences/source-brief.mddoc6,588 bytesreferences/spreadsheet-source.mddoc14,401 bytesreferences/ui-archetypes.mddoc19,180 bytesreferences/ui-direction.mddoc19,431 bytesreferences/ui-palettes.mddoc23,241 bytesreferences/where-it-runs.mddoc4,905 bytes
Read SKILL.md
# Turn Into App
Give a proven workflow an Agent-Native app face and agent brain. Start with
populated domain screens. Object controls send work to the agent; the object
shows progress and receives the result. Avoid form-under-stepper layouts.
## Host, source, and when to ask
### Classify the host
- Local scaffolding needs Node.js 22.22.0 or newer (with npm/npx) and pnpm.
- **Local host**: a terminal, a filesystem, and a target checkout (Claude Code,
Codex, Cursor, or an Agent-Native agent that can edit files). Build the app in
that checkout with steps 1-6, unless the route is Builder hosting (below).View the complete source on GitHubturn-into-skill
When your agent uses it
Capture a proven thread or workflow as a reusable Agent-Native skill and save it as a personal or organization resource. Use when a user invokes `/turn-into-skill` or asks to save a workflow for reuse.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
1,147 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Default behavior
- 1. Extract the repeatable workflow
- 2. Author the skill
- 3. Save at the requested scope
- 4. Verify and hand off
- Related skills
Files
SKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesagents/openai.yamlasset316 bytes
Read SKILL.md
# Turn Into Skill
This is a packaging-and-save workflow, not a request for a proposal. A skill
is the lightweight path for a repeatable job that mostly needs guidance and
existing app actions. Use `turn-into-app` when the job needs a dedicated UI,
new persistent domain state, or a distinct app route.
## Default behavior
- With no argument, use the current thread as the source.
- With a named skill, local workflow, transcript path, or attachment, read that
source and package it immediately.
- Do not ask the user to restate context already present in the source.
- Default to a personal skill unless the user explicitly asks for a team,View the complete source on GitHubworkspace-conventions
When your agent uses it
Shared conventions for a multi-app Agent-Native workspace: finding version-matched framework docs and source, shared vs app-owned code, file and blob storage, env and secrets, agent scratch files, and Dispatch Resources. Use when working across workspace apps, storing files or credentials, or looking up framework APIs.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
1,377 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Framework Docs Lookup
- Shared Conventions
- Related Skills
Files
SKILL.mdcore6,092 bytes
Read SKILL.md
# Workspace Conventions
The long-form detail behind the workspace `AGENTS.md` core rules. Read this
before adding shared code, storing files or secrets, creating workspace
resources, or looking up framework APIs.
## Framework Docs Lookup
Read `agent-native-docs` for the lookup workflow (`framework-search` first, then
the focused `docs-search` / `source-search` readers, and the `rg` fallback).
The workspace-specific slugs it does not list are `workspace` and
`multi-app-workspace`. Feature flags and other specialized workflows are
optional; read their skill only when the app uses them. Use package docs for framework APIs,View the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubexternal-agents
When your agent uses it
Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.
What it loads into context
Frontmatter
148 tokensAlways available to the agent
SKILL.md
6,857 tokensWhen the skill is activated
Reference documents
2,355 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- App-facing MCP connectors
- Provider API connectors
- How
- Do
- Don't
- Related Skills
- Blueprint installer
Files
SKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesreferences/mcp-apps-embedding.mddoc10,363 bytes
Read SKILL.md
# External Agents (MCP bridge + deep links)
## Rule
- **The connected model is the author.** Write content yourself with create/update tools; never delegate via `ask_app` or wait on an unreachable in-app form — see `DEFAULT_AGENT_NATIVE_MCP_INSTRUCTIONS`.
- **Result shape and key tools are in `actions`** (Return Values, Key Actions).
- **Artifacts fit in one call.** Up to 500,000 characters; return a `storing-data` handle.
- **One tab, one id.** WebMCP sends `X-Agent-Native-Browser-Tab`; state resolves to the driving tab (`context-awareness`).
An agent-native app is reachable by any MCP-compatible host. Keep
setup simple: for workspace or cross-app access, add one remote MCP connector:
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page
controls whether that single connector reaches all apps or only selected apps,
and Dispatch filters `list_apps`, `ask_app`, and `open_app` to the granted set.View the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubportability
When your agent uses it
How to keep template code PostgreSQL-specific and hosting-agnostic. Use when defining schemas, writing raw SQL, or creating server routes.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
956 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Database
- Hosting Agnostic
- Related Skills
Files
SKILL.mdcore4,257 bytesSKILL.mdcore4,257 bytes
Read SKILL.md
# Portability
## Rule
Templates use one PostgreSQL schema and query contract. Local PGlite and hosted
PostgreSQL share the same SQL semantics.
## Database
Use Drizzle's PostgreSQL exports directly for schemas and its query builder for reads/writes:
```ts
import { sql } from "drizzle-orm";
import {View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubturn-into-app
When your agent uses it
Turns a thread, skill, spreadsheet, or Claude/ChatGPT project into a polished, visual Agent-Native app: populated domain screens with the in-app agent working behind contextual controls. Use when a user invokes `/turn-into-app`, or asks to turn a workflow, skill, spreadsheet, or project into an app, UI, workbench, or dashboard, including from Claude or ChatGPT on the web.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
3,995 tokensWhen the skill is activated
Reference documents
33,236 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Host, source, and when to ask
- Workflow
- Final report
- Related skills
Files
SKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesagents/openai.yamlasset383 bytesreferences/attachments.mddoc1,503 bytesreferences/fresh-project.mddoc4,733 bytesreferences/local-run-and-deploy.mddoc7,649 bytesreferences/review-loop.mddoc19,749 bytesreferences/source-brief.mddoc6,588 bytesreferences/spreadsheet-source.mddoc14,401 bytesreferences/ui-archetypes.mddoc19,180 bytesreferences/ui-direction.mddoc19,431 bytesreferences/ui-palettes.mddoc23,241 bytesreferences/where-it-runs.mddoc4,905 bytes
Read SKILL.md
# Turn Into App
Give a proven workflow an Agent-Native app face and agent brain. Start with
populated domain screens. Object controls send work to the agent; the object
shows progress and receives the result. Avoid form-under-stepper layouts.
## Host, source, and when to ask
### Classify the host
- Local scaffolding needs Node.js 22.22.0 or newer (with npm/npx) and pnpm.
- **Local host**: a terminal, a filesystem, and a target checkout (Claude Code,
Codex, Cursor, or an Agent-Native agent that can edit files). Build the app in
that checkout with steps 1-6, unless the route is Builder hosting (below).View the complete source on GitHubturn-into-skill
When your agent uses it
Capture a proven thread or workflow as a reusable Agent-Native skill and save it as a personal or organization resource. Use when a user invokes `/turn-into-skill` or asks to save a workflow for reuse.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
1,147 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Default behavior
- 1. Extract the repeatable workflow
- 2. Author the skill
- 3. Save at the requested scope
- 4. Verify and hand off
- Related skills
Files
SKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesagents/openai.yamlasset316 bytes
Read SKILL.md
# Turn Into Skill
This is a packaging-and-save workflow, not a request for a proposal. A skill
is the lightweight path for a repeatable job that mostly needs guidance and
existing app actions. Use `turn-into-app` when the job needs a dedicated UI,
new persistent domain state, or a distinct app route.
## Default behavior
- With no argument, use the current thread as the source.
- With a named skill, local workflow, transcript path, or attachment, read that
source and package it immediately.
- Do not ask the user to restate context already present in the source.
- Default to a personal skill unless the user explicitly asks for a team,View the complete source on GitHubworkspace-conventions
When your agent uses it
Shared conventions for a multi-app Agent-Native workspace: finding version-matched framework docs and source, shared vs app-owned code, file and blob storage, env and secrets, agent scratch files, and Dispatch Resources. Use when working across workspace apps, storing files or credentials, or looking up framework APIs.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
1,377 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Framework Docs Lookup
- Shared Conventions
- Related Skills
Files
SKILL.mdcore6,092 bytes
Read SKILL.md
# Workspace Conventions
The long-form detail behind the workspace `AGENTS.md` core rules. Read this
before adding shared code, storing files or secrets, creating workspace
resources, or looking up framework APIs.
## Framework Docs Lookup
Read `agent-native-docs` for the lookup workflow (`framework-search` first, then
the focused `docs-search` / `source-search` readers, and the `rg` fallback).
The workspace-specific slugs it does not list are `workspace` and
`multi-app-workspace`. Feature flags and other specialized workflows are
optional; read their skill only when the app uses them. Use package docs for framework APIs,View the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubexternal-agents
When your agent uses it
Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.
What it loads into context
Frontmatter
148 tokensAlways available to the agent
SKILL.md
6,857 tokensWhen the skill is activated
Reference documents
2,355 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- App-facing MCP connectors
- Provider API connectors
- How
- Do
- Don't
- Related Skills
- Blueprint installer
Files
SKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesreferences/mcp-apps-embedding.mddoc10,363 bytes
Read SKILL.md
# External Agents (MCP bridge + deep links)
## Rule
- **The connected model is the author.** Write content yourself with create/update tools; never delegate via `ask_app` or wait on an unreachable in-app form — see `DEFAULT_AGENT_NATIVE_MCP_INSTRUCTIONS`.
- **Result shape and key tools are in `actions`** (Return Values, Key Actions).
- **Artifacts fit in one call.** Up to 500,000 characters; return a `storing-data` handle.
- **One tab, one id.** WebMCP sends `X-Agent-Native-Browser-Tab`; state resolves to the driving tab (`context-awareness`).
An agent-native app is reachable by any MCP-compatible host. Keep
setup simple: for workspace or cross-app access, add one remote MCP connector:
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page
controls whether that single connector reaches all apps or only selected apps,
and Dispatch filters `list_apps`, `ask_app`, and `open_app` to the granted set.View the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubportability
When your agent uses it
How to keep template code PostgreSQL-specific and hosting-agnostic. Use when defining schemas, writing raw SQL, or creating server routes.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
956 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Database
- Hosting Agnostic
- Related Skills
Files
SKILL.mdcore4,257 bytesSKILL.mdcore4,257 bytes
Read SKILL.md
# Portability
## Rule
Templates use one PostgreSQL schema and query contract. Local PGlite and hosted
PostgreSQL share the same SQL semantics.
## Database
Use Drizzle's PostgreSQL exports directly for schemas and its query builder for reads/writes:
```ts
import { sql } from "drizzle-orm";
import {View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubturn-into-app
When your agent uses it
Turns a thread, skill, spreadsheet, or Claude/ChatGPT project into a polished, visual Agent-Native app: populated domain screens with the in-app agent working behind contextual controls. Use when a user invokes `/turn-into-app`, or asks to turn a workflow, skill, spreadsheet, or project into an app, UI, workbench, or dashboard, including from Claude or ChatGPT on the web.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
3,995 tokensWhen the skill is activated
Reference documents
33,236 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Host, source, and when to ask
- Workflow
- Final report
- Related skills
Files
SKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesagents/openai.yamlasset383 bytesreferences/attachments.mddoc1,503 bytesreferences/fresh-project.mddoc4,733 bytesreferences/local-run-and-deploy.mddoc7,649 bytesreferences/review-loop.mddoc19,749 bytesreferences/source-brief.mddoc6,588 bytesreferences/spreadsheet-source.mddoc14,401 bytesreferences/ui-archetypes.mddoc19,180 bytesreferences/ui-direction.mddoc19,431 bytesreferences/ui-palettes.mddoc23,241 bytesreferences/where-it-runs.mddoc4,905 bytes
Read SKILL.md
# Turn Into App
Give a proven workflow an Agent-Native app face and agent brain. Start with
populated domain screens. Object controls send work to the agent; the object
shows progress and receives the result. Avoid form-under-stepper layouts.
## Host, source, and when to ask
### Classify the host
- Local scaffolding needs Node.js 22.22.0 or newer (with npm/npx) and pnpm.
- **Local host**: a terminal, a filesystem, and a target checkout (Claude Code,
Codex, Cursor, or an Agent-Native agent that can edit files). Build the app in
that checkout with steps 1-6, unless the route is Builder hosting (below).View the complete source on GitHubturn-into-skill
When your agent uses it
Capture a proven thread or workflow as a reusable Agent-Native skill and save it as a personal or organization resource. Use when a user invokes `/turn-into-skill` or asks to save a workflow for reuse.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
1,147 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Default behavior
- 1. Extract the repeatable workflow
- 2. Author the skill
- 3. Save at the requested scope
- 4. Verify and hand off
- Related skills
Files
SKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesagents/openai.yamlasset316 bytes
Read SKILL.md
# Turn Into Skill
This is a packaging-and-save workflow, not a request for a proposal. A skill
is the lightweight path for a repeatable job that mostly needs guidance and
existing app actions. Use `turn-into-app` when the job needs a dedicated UI,
new persistent domain state, or a distinct app route.
## Default behavior
- With no argument, use the current thread as the source.
- With a named skill, local workflow, transcript path, or attachment, read that
source and package it immediately.
- Do not ask the user to restate context already present in the source.
- Default to a personal skill unless the user explicitly asks for a team,View the complete source on GitHubworkspace-conventions
When your agent uses it
Shared conventions for a multi-app Agent-Native workspace: finding version-matched framework docs and source, shared vs app-owned code, file and blob storage, env and secrets, agent scratch files, and Dispatch Resources. Use when working across workspace apps, storing files or credentials, or looking up framework APIs.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
1,377 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Framework Docs Lookup
- Shared Conventions
- Related Skills
Files
SKILL.mdcore6,092 bytes
Read SKILL.md
# Workspace Conventions
The long-form detail behind the workspace `AGENTS.md` core rules. Read this
before adding shared code, storing files or secrets, creating workspace
resources, or looking up framework APIs.
## Framework Docs Lookup
Read `agent-native-docs` for the lookup workflow (`framework-search` first, then
the focused `docs-search` / `source-search` readers, and the `rg` fallback).
The workspace-specific slugs it does not list are `workspace` and
`multi-app-workspace`. Feature flags and other specialized workflows are
optional; read their skill only when the app uses them. Use package docs for framework APIs,View the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubexternal-agents
When your agent uses it
Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.
What it loads into context
Frontmatter
148 tokensAlways available to the agent
SKILL.md
6,857 tokensWhen the skill is activated
Reference documents
2,355 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- App-facing MCP connectors
- Provider API connectors
- How
- Do
- Don't
- Related Skills
- Blueprint installer
Files
SKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesreferences/mcp-apps-embedding.mddoc10,363 bytes
Read SKILL.md
# External Agents (MCP bridge + deep links)
## Rule
- **The connected model is the author.** Write content yourself with create/update tools; never delegate via `ask_app` or wait on an unreachable in-app form — see `DEFAULT_AGENT_NATIVE_MCP_INSTRUCTIONS`.
- **Result shape and key tools are in `actions`** (Return Values, Key Actions).
- **Artifacts fit in one call.** Up to 500,000 characters; return a `storing-data` handle.
- **One tab, one id.** WebMCP sends `X-Agent-Native-Browser-Tab`; state resolves to the driving tab (`context-awareness`).
An agent-native app is reachable by any MCP-compatible host. Keep
setup simple: for workspace or cross-app access, add one remote MCP connector:
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page
controls whether that single connector reaches all apps or only selected apps,
and Dispatch filters `list_apps`, `ask_app`, and `open_app` to the granted set.View the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubportability
When your agent uses it
How to keep template code PostgreSQL-specific and hosting-agnostic. Use when defining schemas, writing raw SQL, or creating server routes.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
956 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Database
- Hosting Agnostic
- Related Skills
Files
SKILL.mdcore4,257 bytesSKILL.mdcore4,257 bytes
Read SKILL.md
# Portability
## Rule
Templates use one PostgreSQL schema and query contract. Local PGlite and hosted
PostgreSQL share the same SQL semantics.
## Database
Use Drizzle's PostgreSQL exports directly for schemas and its query builder for reads/writes:
```ts
import { sql } from "drizzle-orm";
import {View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubturn-into-app
When your agent uses it
Turns a thread, skill, spreadsheet, or Claude/ChatGPT project into a polished, visual Agent-Native app: populated domain screens with the in-app agent working behind contextual controls. Use when a user invokes `/turn-into-app`, or asks to turn a workflow, skill, spreadsheet, or project into an app, UI, workbench, or dashboard, including from Claude or ChatGPT on the web.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
3,995 tokensWhen the skill is activated
Reference documents
33,236 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Host, source, and when to ask
- Workflow
- Final report
- Related skills
Files
SKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesagents/openai.yamlasset383 bytesreferences/attachments.mddoc1,503 bytesreferences/fresh-project.mddoc4,733 bytesreferences/local-run-and-deploy.mddoc7,649 bytesreferences/review-loop.mddoc19,749 bytesreferences/source-brief.mddoc6,588 bytesreferences/spreadsheet-source.mddoc14,401 bytesreferences/ui-archetypes.mddoc19,180 bytesreferences/ui-direction.mddoc19,431 bytesreferences/ui-palettes.mddoc23,241 bytesreferences/where-it-runs.mddoc4,905 bytes
Read SKILL.md
# Turn Into App
Give a proven workflow an Agent-Native app face and agent brain. Start with
populated domain screens. Object controls send work to the agent; the object
shows progress and receives the result. Avoid form-under-stepper layouts.
## Host, source, and when to ask
### Classify the host
- Local scaffolding needs Node.js 22.22.0 or newer (with npm/npx) and pnpm.
- **Local host**: a terminal, a filesystem, and a target checkout (Claude Code,
Codex, Cursor, or an Agent-Native agent that can edit files). Build the app in
that checkout with steps 1-6, unless the route is Builder hosting (below).View the complete source on GitHubturn-into-skill
When your agent uses it
Capture a proven thread or workflow as a reusable Agent-Native skill and save it as a personal or organization resource. Use when a user invokes `/turn-into-skill` or asks to save a workflow for reuse.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
1,147 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Default behavior
- 1. Extract the repeatable workflow
- 2. Author the skill
- 3. Save at the requested scope
- 4. Verify and hand off
- Related skills
Files
SKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesagents/openai.yamlasset316 bytes
Read SKILL.md
# Turn Into Skill
This is a packaging-and-save workflow, not a request for a proposal. A skill
is the lightweight path for a repeatable job that mostly needs guidance and
existing app actions. Use `turn-into-app` when the job needs a dedicated UI,
new persistent domain state, or a distinct app route.
## Default behavior
- With no argument, use the current thread as the source.
- With a named skill, local workflow, transcript path, or attachment, read that
source and package it immediately.
- Do not ask the user to restate context already present in the source.
- Default to a personal skill unless the user explicitly asks for a team,View the complete source on GitHubworkspace-conventions
When your agent uses it
Shared conventions for a multi-app Agent-Native workspace: finding version-matched framework docs and source, shared vs app-owned code, file and blob storage, env and secrets, agent scratch files, and Dispatch Resources. Use when working across workspace apps, storing files or credentials, or looking up framework APIs.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
1,377 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Framework Docs Lookup
- Shared Conventions
- Related Skills
Files
SKILL.mdcore6,092 bytes
Read SKILL.md
# Workspace Conventions
The long-form detail behind the workspace `AGENTS.md` core rules. Read this
before adding shared code, storing files or secrets, creating workspace
resources, or looking up framework APIs.
## Framework Docs Lookup
Read `agent-native-docs` for the lookup workflow (`framework-search` first, then
the focused `docs-search` / `source-search` readers, and the `rg` fallback).
The workspace-specific slugs it does not list are `workspace` and
`multi-app-workspace`. Feature flags and other specialized workflows are
optional; read their skill only when the app uses them. Use package docs for framework APIs,View the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubexternal-agents
When your agent uses it
Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.
What it loads into context
Frontmatter
148 tokensAlways available to the agent
SKILL.md
6,857 tokensWhen the skill is activated
Reference documents
2,355 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- App-facing MCP connectors
- Provider API connectors
- How
- Do
- Don't
- Related Skills
- Blueprint installer
Files
SKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesreferences/mcp-apps-embedding.mddoc10,363 bytes
Read SKILL.md
# External Agents (MCP bridge + deep links)
## Rule
- **The connected model is the author.** Write content yourself with create/update tools; never delegate via `ask_app` or wait on an unreachable in-app form — see `DEFAULT_AGENT_NATIVE_MCP_INSTRUCTIONS`.
- **Result shape and key tools are in `actions`** (Return Values, Key Actions).
- **Artifacts fit in one call.** Up to 500,000 characters; return a `storing-data` handle.
- **One tab, one id.** WebMCP sends `X-Agent-Native-Browser-Tab`; state resolves to the driving tab (`context-awareness`).
An agent-native app is reachable by any MCP-compatible host. Keep
setup simple: for workspace or cross-app access, add one remote MCP connector:
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page
controls whether that single connector reaches all apps or only selected apps,
and Dispatch filters `list_apps`, `ask_app`, and `open_app` to the granted set.View the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubportability
When your agent uses it
How to keep template code PostgreSQL-specific and hosting-agnostic. Use when defining schemas, writing raw SQL, or creating server routes.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
956 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Database
- Hosting Agnostic
- Related Skills
Files
SKILL.mdcore4,257 bytesSKILL.mdcore4,257 bytes
Read SKILL.md
# Portability
## Rule
Templates use one PostgreSQL schema and query contract. Local PGlite and hosted
PostgreSQL share the same SQL semantics.
## Database
Use Drizzle's PostgreSQL exports directly for schemas and its query builder for reads/writes:
```ts
import { sql } from "drizzle-orm";
import {View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubturn-into-app
When your agent uses it
Turns a thread, skill, spreadsheet, or Claude/ChatGPT project into a polished, visual Agent-Native app: populated domain screens with the in-app agent working behind contextual controls. Use when a user invokes `/turn-into-app`, or asks to turn a workflow, skill, spreadsheet, or project into an app, UI, workbench, or dashboard, including from Claude or ChatGPT on the web.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
3,995 tokensWhen the skill is activated
Reference documents
33,236 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Host, source, and when to ask
- Workflow
- Final report
- Related skills
Files
SKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesagents/openai.yamlasset383 bytesreferences/attachments.mddoc1,503 bytesreferences/fresh-project.mddoc4,733 bytesreferences/local-run-and-deploy.mddoc7,649 bytesreferences/review-loop.mddoc19,749 bytesreferences/source-brief.mddoc6,588 bytesreferences/spreadsheet-source.mddoc14,401 bytesreferences/ui-archetypes.mddoc19,180 bytesreferences/ui-direction.mddoc19,431 bytesreferences/ui-palettes.mddoc23,241 bytesreferences/where-it-runs.mddoc4,905 bytes
Read SKILL.md
# Turn Into App
Give a proven workflow an Agent-Native app face and agent brain. Start with
populated domain screens. Object controls send work to the agent; the object
shows progress and receives the result. Avoid form-under-stepper layouts.
## Host, source, and when to ask
### Classify the host
- Local scaffolding needs Node.js 22.22.0 or newer (with npm/npx) and pnpm.
- **Local host**: a terminal, a filesystem, and a target checkout (Claude Code,
Codex, Cursor, or an Agent-Native agent that can edit files). Build the app in
that checkout with steps 1-6, unless the route is Builder hosting (below).View the complete source on GitHubturn-into-skill
When your agent uses it
Capture a proven thread or workflow as a reusable Agent-Native skill and save it as a personal or organization resource. Use when a user invokes `/turn-into-skill` or asks to save a workflow for reuse.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
1,147 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Default behavior
- 1. Extract the repeatable workflow
- 2. Author the skill
- 3. Save at the requested scope
- 4. Verify and hand off
- Related skills
Files
SKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesagents/openai.yamlasset316 bytes
Read SKILL.md
# Turn Into Skill
This is a packaging-and-save workflow, not a request for a proposal. A skill
is the lightweight path for a repeatable job that mostly needs guidance and
existing app actions. Use `turn-into-app` when the job needs a dedicated UI,
new persistent domain state, or a distinct app route.
## Default behavior
- With no argument, use the current thread as the source.
- With a named skill, local workflow, transcript path, or attachment, read that
source and package it immediately.
- Do not ask the user to restate context already present in the source.
- Default to a personal skill unless the user explicitly asks for a team,View the complete source on GitHubworkspace-conventions
When your agent uses it
Shared conventions for a multi-app Agent-Native workspace: finding version-matched framework docs and source, shared vs app-owned code, file and blob storage, env and secrets, agent scratch files, and Dispatch Resources. Use when working across workspace apps, storing files or credentials, or looking up framework APIs.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
1,377 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Framework Docs Lookup
- Shared Conventions
- Related Skills
Files
SKILL.mdcore6,092 bytes
Read SKILL.md
# Workspace Conventions
The long-form detail behind the workspace `AGENTS.md` core rules. Read this
before adding shared code, storing files or secrets, creating workspace
resources, or looking up framework APIs.
## Framework Docs Lookup
Read `agent-native-docs` for the lookup workflow (`framework-search` first, then
the focused `docs-search` / `source-search` readers, and the `rg` fallback).
The workspace-specific slugs it does not list are `workspace` and
`multi-app-workspace`. Feature flags and other specialized workflows are
optional; read their skill only when the app uses them. Use package docs for framework APIs,View the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubexternal-agents
When your agent uses it
Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.
What it loads into context
Frontmatter
148 tokensAlways available to the agent
SKILL.md
6,857 tokensWhen the skill is activated
Reference documents
2,355 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- App-facing MCP connectors
- Provider API connectors
- How
- Do
- Don't
- Related Skills
- Blueprint installer
Files
SKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesreferences/mcp-apps-embedding.mddoc10,363 bytes
Read SKILL.md
# External Agents (MCP bridge + deep links)
## Rule
- **The connected model is the author.** Write content yourself with create/update tools; never delegate via `ask_app` or wait on an unreachable in-app form — see `DEFAULT_AGENT_NATIVE_MCP_INSTRUCTIONS`.
- **Result shape and key tools are in `actions`** (Return Values, Key Actions).
- **Artifacts fit in one call.** Up to 500,000 characters; return a `storing-data` handle.
- **One tab, one id.** WebMCP sends `X-Agent-Native-Browser-Tab`; state resolves to the driving tab (`context-awareness`).
An agent-native app is reachable by any MCP-compatible host. Keep
setup simple: for workspace or cross-app access, add one remote MCP connector:
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page
controls whether that single connector reaches all apps or only selected apps,
and Dispatch filters `list_apps`, `ask_app`, and `open_app` to the granted set.View the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubportability
When your agent uses it
How to keep template code PostgreSQL-specific and hosting-agnostic. Use when defining schemas, writing raw SQL, or creating server routes.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
956 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Database
- Hosting Agnostic
- Related Skills
Files
SKILL.mdcore4,257 bytesSKILL.mdcore4,257 bytes
Read SKILL.md
# Portability
## Rule
Templates use one PostgreSQL schema and query contract. Local PGlite and hosted
PostgreSQL share the same SQL semantics.
## Database
Use Drizzle's PostgreSQL exports directly for schemas and its query builder for reads/writes:
```ts
import { sql } from "drizzle-orm";
import {View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubturn-into-app
When your agent uses it
Turns a thread, skill, spreadsheet, or Claude/ChatGPT project into a polished, visual Agent-Native app: populated domain screens with the in-app agent working behind contextual controls. Use when a user invokes `/turn-into-app`, or asks to turn a workflow, skill, spreadsheet, or project into an app, UI, workbench, or dashboard, including from Claude or ChatGPT on the web.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
3,995 tokensWhen the skill is activated
Reference documents
33,236 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Host, source, and when to ask
- Workflow
- Final report
- Related skills
Files
SKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesagents/openai.yamlasset383 bytesreferences/attachments.mddoc1,503 bytesreferences/fresh-project.mddoc4,733 bytesreferences/local-run-and-deploy.mddoc7,649 bytesreferences/review-loop.mddoc19,749 bytesreferences/source-brief.mddoc6,588 bytesreferences/spreadsheet-source.mddoc14,401 bytesreferences/ui-archetypes.mddoc19,180 bytesreferences/ui-direction.mddoc19,431 bytesreferences/ui-palettes.mddoc23,241 bytesreferences/where-it-runs.mddoc4,905 bytes
Read SKILL.md
# Turn Into App
Give a proven workflow an Agent-Native app face and agent brain. Start with
populated domain screens. Object controls send work to the agent; the object
shows progress and receives the result. Avoid form-under-stepper layouts.
## Host, source, and when to ask
### Classify the host
- Local scaffolding needs Node.js 22.22.0 or newer (with npm/npx) and pnpm.
- **Local host**: a terminal, a filesystem, and a target checkout (Claude Code,
Codex, Cursor, or an Agent-Native agent that can edit files). Build the app in
that checkout with steps 1-6, unless the route is Builder hosting (below).View the complete source on GitHubturn-into-skill
When your agent uses it
Capture a proven thread or workflow as a reusable Agent-Native skill and save it as a personal or organization resource. Use when a user invokes `/turn-into-skill` or asks to save a workflow for reuse.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
1,147 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Default behavior
- 1. Extract the repeatable workflow
- 2. Author the skill
- 3. Save at the requested scope
- 4. Verify and hand off
- Related skills
Files
SKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesagents/openai.yamlasset316 bytes
Read SKILL.md
# Turn Into Skill
This is a packaging-and-save workflow, not a request for a proposal. A skill
is the lightweight path for a repeatable job that mostly needs guidance and
existing app actions. Use `turn-into-app` when the job needs a dedicated UI,
new persistent domain state, or a distinct app route.
## Default behavior
- With no argument, use the current thread as the source.
- With a named skill, local workflow, transcript path, or attachment, read that
source and package it immediately.
- Do not ask the user to restate context already present in the source.
- Default to a personal skill unless the user explicitly asks for a team,View the complete source on GitHubworkspace-conventions
When your agent uses it
Shared conventions for a multi-app Agent-Native workspace: finding version-matched framework docs and source, shared vs app-owned code, file and blob storage, env and secrets, agent scratch files, and Dispatch Resources. Use when working across workspace apps, storing files or credentials, or looking up framework APIs.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
1,377 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Framework Docs Lookup
- Shared Conventions
- Related Skills
Files
SKILL.mdcore6,092 bytes
Read SKILL.md
# Workspace Conventions
The long-form detail behind the workspace `AGENTS.md` core rules. Read this
before adding shared code, storing files or secrets, creating workspace
resources, or looking up framework APIs.
## Framework Docs Lookup
Read `agent-native-docs` for the lookup workflow (`framework-search` first, then
the focused `docs-search` / `source-search` readers, and the `rg` fallback).
The workspace-specific slugs it does not list are `workspace` and
`multi-app-workspace`. Feature flags and other specialized workflows are
optional; read their skill only when the app uses them. Use package docs for framework APIs,View the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubexternal-agents
When your agent uses it
Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.
What it loads into context
Frontmatter
148 tokensAlways available to the agent
SKILL.md
6,857 tokensWhen the skill is activated
Reference documents
2,355 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- App-facing MCP connectors
- Provider API connectors
- How
- Do
- Don't
- Related Skills
- Blueprint installer
Files
SKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesreferences/mcp-apps-embedding.mddoc10,363 bytes
Read SKILL.md
# External Agents (MCP bridge + deep links)
## Rule
- **The connected model is the author.** Write content yourself with create/update tools; never delegate via `ask_app` or wait on an unreachable in-app form — see `DEFAULT_AGENT_NATIVE_MCP_INSTRUCTIONS`.
- **Result shape and key tools are in `actions`** (Return Values, Key Actions).
- **Artifacts fit in one call.** Up to 500,000 characters; return a `storing-data` handle.
- **One tab, one id.** WebMCP sends `X-Agent-Native-Browser-Tab`; state resolves to the driving tab (`context-awareness`).
An agent-native app is reachable by any MCP-compatible host. Keep
setup simple: for workspace or cross-app access, add one remote MCP connector:
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page
controls whether that single connector reaches all apps or only selected apps,
and Dispatch filters `list_apps`, `ask_app`, and `open_app` to the granted set.View the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubportability
When your agent uses it
How to keep template code PostgreSQL-specific and hosting-agnostic. Use when defining schemas, writing raw SQL, or creating server routes.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
956 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Database
- Hosting Agnostic
- Related Skills
Files
SKILL.mdcore4,257 bytesSKILL.mdcore4,257 bytes
Read SKILL.md
# Portability
## Rule
Templates use one PostgreSQL schema and query contract. Local PGlite and hosted
PostgreSQL share the same SQL semantics.
## Database
Use Drizzle's PostgreSQL exports directly for schemas and its query builder for reads/writes:
```ts
import { sql } from "drizzle-orm";
import {View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubturn-into-app
When your agent uses it
Turns a thread, skill, spreadsheet, or Claude/ChatGPT project into a polished, visual Agent-Native app: populated domain screens with the in-app agent working behind contextual controls. Use when a user invokes `/turn-into-app`, or asks to turn a workflow, skill, spreadsheet, or project into an app, UI, workbench, or dashboard, including from Claude or ChatGPT on the web.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
3,995 tokensWhen the skill is activated
Reference documents
33,236 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Host, source, and when to ask
- Workflow
- Final report
- Related skills
Files
SKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesagents/openai.yamlasset383 bytesreferences/attachments.mddoc1,503 bytesreferences/fresh-project.mddoc4,733 bytesreferences/local-run-and-deploy.mddoc7,649 bytesreferences/review-loop.mddoc19,749 bytesreferences/source-brief.mddoc6,588 bytesreferences/spreadsheet-source.mddoc14,401 bytesreferences/ui-archetypes.mddoc19,180 bytesreferences/ui-direction.mddoc19,431 bytesreferences/ui-palettes.mddoc23,241 bytesreferences/where-it-runs.mddoc4,905 bytes
Read SKILL.md
# Turn Into App
Give a proven workflow an Agent-Native app face and agent brain. Start with
populated domain screens. Object controls send work to the agent; the object
shows progress and receives the result. Avoid form-under-stepper layouts.
## Host, source, and when to ask
### Classify the host
- Local scaffolding needs Node.js 22.22.0 or newer (with npm/npx) and pnpm.
- **Local host**: a terminal, a filesystem, and a target checkout (Claude Code,
Codex, Cursor, or an Agent-Native agent that can edit files). Build the app in
that checkout with steps 1-6, unless the route is Builder hosting (below).View the complete source on GitHubturn-into-skill
When your agent uses it
Capture a proven thread or workflow as a reusable Agent-Native skill and save it as a personal or organization resource. Use when a user invokes `/turn-into-skill` or asks to save a workflow for reuse.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
1,147 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Default behavior
- 1. Extract the repeatable workflow
- 2. Author the skill
- 3. Save at the requested scope
- 4. Verify and hand off
- Related skills
Files
SKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesagents/openai.yamlasset316 bytes
Read SKILL.md
# Turn Into Skill
This is a packaging-and-save workflow, not a request for a proposal. A skill
is the lightweight path for a repeatable job that mostly needs guidance and
existing app actions. Use `turn-into-app` when the job needs a dedicated UI,
new persistent domain state, or a distinct app route.
## Default behavior
- With no argument, use the current thread as the source.
- With a named skill, local workflow, transcript path, or attachment, read that
source and package it immediately.
- Do not ask the user to restate context already present in the source.
- Default to a personal skill unless the user explicitly asks for a team,View the complete source on GitHubworkspace-conventions
When your agent uses it
Shared conventions for a multi-app Agent-Native workspace: finding version-matched framework docs and source, shared vs app-owned code, file and blob storage, env and secrets, agent scratch files, and Dispatch Resources. Use when working across workspace apps, storing files or credentials, or looking up framework APIs.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
1,377 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Framework Docs Lookup
- Shared Conventions
- Related Skills
Files
SKILL.mdcore6,092 bytes
Read SKILL.md
# Workspace Conventions
The long-form detail behind the workspace `AGENTS.md` core rules. Read this
before adding shared code, storing files or secrets, creating workspace
resources, or looking up framework APIs.
## Framework Docs Lookup
Read `agent-native-docs` for the lookup workflow (`framework-search` first, then
the focused `docs-search` / `source-search` readers, and the `rg` fallback).
The workspace-specific slugs it does not list are `workspace` and
`multi-app-workspace`. Feature flags and other specialized workflows are
optional; read their skill only when the app uses them. Use package docs for framework APIs,View the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubexternal-agents
When your agent uses it
Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.
What it loads into context
Frontmatter
148 tokensAlways available to the agent
SKILL.md
6,857 tokensWhen the skill is activated
Reference documents
2,355 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- App-facing MCP connectors
- Provider API connectors
- How
- Do
- Don't
- Related Skills
- Blueprint installer
Files
SKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesreferences/mcp-apps-embedding.mddoc10,363 bytes
Read SKILL.md
# External Agents (MCP bridge + deep links)
## Rule
- **The connected model is the author.** Write content yourself with create/update tools; never delegate via `ask_app` or wait on an unreachable in-app form — see `DEFAULT_AGENT_NATIVE_MCP_INSTRUCTIONS`.
- **Result shape and key tools are in `actions`** (Return Values, Key Actions).
- **Artifacts fit in one call.** Up to 500,000 characters; return a `storing-data` handle.
- **One tab, one id.** WebMCP sends `X-Agent-Native-Browser-Tab`; state resolves to the driving tab (`context-awareness`).
An agent-native app is reachable by any MCP-compatible host. Keep
setup simple: for workspace or cross-app access, add one remote MCP connector:
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page
controls whether that single connector reaches all apps or only selected apps,
and Dispatch filters `list_apps`, `ask_app`, and `open_app` to the granted set.View the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubportability
When your agent uses it
How to keep template code PostgreSQL-specific and hosting-agnostic. Use when defining schemas, writing raw SQL, or creating server routes.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
956 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Database
- Hosting Agnostic
- Related Skills
Files
SKILL.mdcore4,257 bytesSKILL.mdcore4,257 bytes
Read SKILL.md
# Portability
## Rule
Templates use one PostgreSQL schema and query contract. Local PGlite and hosted
PostgreSQL share the same SQL semantics.
## Database
Use Drizzle's PostgreSQL exports directly for schemas and its query builder for reads/writes:
```ts
import { sql } from "drizzle-orm";
import {View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubturn-into-app
When your agent uses it
Turns a thread, skill, spreadsheet, or Claude/ChatGPT project into a polished, visual Agent-Native app: populated domain screens with the in-app agent working behind contextual controls. Use when a user invokes `/turn-into-app`, or asks to turn a workflow, skill, spreadsheet, or project into an app, UI, workbench, or dashboard, including from Claude or ChatGPT on the web.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
3,995 tokensWhen the skill is activated
Reference documents
33,236 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Host, source, and when to ask
- Workflow
- Final report
- Related skills
Files
SKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesagents/openai.yamlasset383 bytesreferences/attachments.mddoc1,503 bytesreferences/fresh-project.mddoc4,733 bytesreferences/local-run-and-deploy.mddoc7,649 bytesreferences/review-loop.mddoc19,749 bytesreferences/source-brief.mddoc6,588 bytesreferences/spreadsheet-source.mddoc14,401 bytesreferences/ui-archetypes.mddoc19,180 bytesreferences/ui-direction.mddoc19,431 bytesreferences/ui-palettes.mddoc23,241 bytesreferences/where-it-runs.mddoc4,905 bytes
Read SKILL.md
# Turn Into App
Give a proven workflow an Agent-Native app face and agent brain. Start with
populated domain screens. Object controls send work to the agent; the object
shows progress and receives the result. Avoid form-under-stepper layouts.
## Host, source, and when to ask
### Classify the host
- Local scaffolding needs Node.js 22.22.0 or newer (with npm/npx) and pnpm.
- **Local host**: a terminal, a filesystem, and a target checkout (Claude Code,
Codex, Cursor, or an Agent-Native agent that can edit files). Build the app in
that checkout with steps 1-6, unless the route is Builder hosting (below).View the complete source on GitHubturn-into-skill
When your agent uses it
Capture a proven thread or workflow as a reusable Agent-Native skill and save it as a personal or organization resource. Use when a user invokes `/turn-into-skill` or asks to save a workflow for reuse.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
1,147 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Default behavior
- 1. Extract the repeatable workflow
- 2. Author the skill
- 3. Save at the requested scope
- 4. Verify and hand off
- Related skills
Files
SKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesagents/openai.yamlasset316 bytes
Read SKILL.md
# Turn Into Skill
This is a packaging-and-save workflow, not a request for a proposal. A skill
is the lightweight path for a repeatable job that mostly needs guidance and
existing app actions. Use `turn-into-app` when the job needs a dedicated UI,
new persistent domain state, or a distinct app route.
## Default behavior
- With no argument, use the current thread as the source.
- With a named skill, local workflow, transcript path, or attachment, read that
source and package it immediately.
- Do not ask the user to restate context already present in the source.
- Default to a personal skill unless the user explicitly asks for a team,View the complete source on GitHubworkspace-conventions
When your agent uses it
Shared conventions for a multi-app Agent-Native workspace: finding version-matched framework docs and source, shared vs app-owned code, file and blob storage, env and secrets, agent scratch files, and Dispatch Resources. Use when working across workspace apps, storing files or credentials, or looking up framework APIs.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
1,377 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Framework Docs Lookup
- Shared Conventions
- Related Skills
Files
SKILL.mdcore6,092 bytes
Read SKILL.md
# Workspace Conventions
The long-form detail behind the workspace `AGENTS.md` core rules. Read this
before adding shared code, storing files or secrets, creating workspace
resources, or looking up framework APIs.
## Framework Docs Lookup
Read `agent-native-docs` for the lookup workflow (`framework-search` first, then
the focused `docs-search` / `source-search` readers, and the `rg` fallback).
The workspace-specific slugs it does not list are `workspace` and
`multi-app-workspace`. Feature flags and other specialized workflows are
optional; read their skill only when the app uses them. Use package docs for framework APIs,View the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubexternal-agents
When your agent uses it
Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.
What it loads into context
Frontmatter
148 tokensAlways available to the agent
SKILL.md
6,857 tokensWhen the skill is activated
Reference documents
2,355 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- App-facing MCP connectors
- Provider API connectors
- How
- Do
- Don't
- Related Skills
- Blueprint installer
Files
SKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesreferences/mcp-apps-embedding.mddoc10,363 bytes
Read SKILL.md
# External Agents (MCP bridge + deep links)
## Rule
- **The connected model is the author.** Write content yourself with create/update tools; never delegate via `ask_app` or wait on an unreachable in-app form — see `DEFAULT_AGENT_NATIVE_MCP_INSTRUCTIONS`.
- **Result shape and key tools are in `actions`** (Return Values, Key Actions).
- **Artifacts fit in one call.** Up to 500,000 characters; return a `storing-data` handle.
- **One tab, one id.** WebMCP sends `X-Agent-Native-Browser-Tab`; state resolves to the driving tab (`context-awareness`).
An agent-native app is reachable by any MCP-compatible host. Keep
setup simple: for workspace or cross-app access, add one remote MCP connector:
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page
controls whether that single connector reaches all apps or only selected apps,
and Dispatch filters `list_apps`, `ask_app`, and `open_app` to the granted set.View the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubportability
When your agent uses it
How to keep template code PostgreSQL-specific and hosting-agnostic. Use when defining schemas, writing raw SQL, or creating server routes.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
956 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Database
- Hosting Agnostic
- Related Skills
Files
SKILL.mdcore4,257 bytesSKILL.mdcore4,257 bytes
Read SKILL.md
# Portability
## Rule
Templates use one PostgreSQL schema and query contract. Local PGlite and hosted
PostgreSQL share the same SQL semantics.
## Database
Use Drizzle's PostgreSQL exports directly for schemas and its query builder for reads/writes:
```ts
import { sql } from "drizzle-orm";
import {View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubturn-into-app
When your agent uses it
Turns a thread, skill, spreadsheet, or Claude/ChatGPT project into a polished, visual Agent-Native app: populated domain screens with the in-app agent working behind contextual controls. Use when a user invokes `/turn-into-app`, or asks to turn a workflow, skill, spreadsheet, or project into an app, UI, workbench, or dashboard, including from Claude or ChatGPT on the web.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
3,995 tokensWhen the skill is activated
Reference documents
33,236 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Host, source, and when to ask
- Workflow
- Final report
- Related skills
Files
SKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesagents/openai.yamlasset383 bytesreferences/attachments.mddoc1,503 bytesreferences/fresh-project.mddoc4,733 bytesreferences/local-run-and-deploy.mddoc7,649 bytesreferences/review-loop.mddoc19,749 bytesreferences/source-brief.mddoc6,588 bytesreferences/spreadsheet-source.mddoc14,401 bytesreferences/ui-archetypes.mddoc19,180 bytesreferences/ui-direction.mddoc19,431 bytesreferences/ui-palettes.mddoc23,241 bytesreferences/where-it-runs.mddoc4,905 bytes
Read SKILL.md
# Turn Into App
Give a proven workflow an Agent-Native app face and agent brain. Start with
populated domain screens. Object controls send work to the agent; the object
shows progress and receives the result. Avoid form-under-stepper layouts.
## Host, source, and when to ask
### Classify the host
- Local scaffolding needs Node.js 22.22.0 or newer (with npm/npx) and pnpm.
- **Local host**: a terminal, a filesystem, and a target checkout (Claude Code,
Codex, Cursor, or an Agent-Native agent that can edit files). Build the app in
that checkout with steps 1-6, unless the route is Builder hosting (below).View the complete source on GitHubturn-into-skill
When your agent uses it
Capture a proven thread or workflow as a reusable Agent-Native skill and save it as a personal or organization resource. Use when a user invokes `/turn-into-skill` or asks to save a workflow for reuse.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
1,147 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Default behavior
- 1. Extract the repeatable workflow
- 2. Author the skill
- 3. Save at the requested scope
- 4. Verify and hand off
- Related skills
Files
SKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesagents/openai.yamlasset316 bytes
Read SKILL.md
# Turn Into Skill
This is a packaging-and-save workflow, not a request for a proposal. A skill
is the lightweight path for a repeatable job that mostly needs guidance and
existing app actions. Use `turn-into-app` when the job needs a dedicated UI,
new persistent domain state, or a distinct app route.
## Default behavior
- With no argument, use the current thread as the source.
- With a named skill, local workflow, transcript path, or attachment, read that
source and package it immediately.
- Do not ask the user to restate context already present in the source.
- Default to a personal skill unless the user explicitly asks for a team,View the complete source on GitHubworkspace-conventions
When your agent uses it
Shared conventions for a multi-app Agent-Native workspace: finding version-matched framework docs and source, shared vs app-owned code, file and blob storage, env and secrets, agent scratch files, and Dispatch Resources. Use when working across workspace apps, storing files or credentials, or looking up framework APIs.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
1,377 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Framework Docs Lookup
- Shared Conventions
- Related Skills
Files
SKILL.mdcore6,092 bytes
Read SKILL.md
# Workspace Conventions
The long-form detail behind the workspace `AGENTS.md` core rules. Read this
before adding shared code, storing files or secrets, creating workspace
resources, or looking up framework APIs.
## Framework Docs Lookup
Read `agent-native-docs` for the lookup workflow (`framework-search` first, then
the focused `docs-search` / `source-search` readers, and the `rg` fallback).
The workspace-specific slugs it does not list are `workspace` and
`multi-app-workspace`. Feature flags and other specialized workflows are
optional; read their skill only when the app uses them. Use package docs for framework APIs,View the complete source on GitHubagent-native-docs
When your agent uses it
How to look up version-matched Agent-Native framework docs and source in node_modules. Use before coding against @agent-native/core APIs or advanced features.
What it loads into context
Frontmatter
41 tokensAlways available to the agent
SKILL.md
550 tokensWhen the skill is activated
Reference documents
27,704 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How
- Useful Slugs
- Don't
Files
AGENTS.mddoc5,945 bytesSKILL.mdcore2,266 bytesSKILL.mdcore2,266 bytesSKILL.mdcore2,266 bytesSKILL.mdcore2,266 bytesSKILL.mdcore2,266 bytesSKILL.mdcore2,266 bytescontent/a2a-protocol.mdxasset38,488 bytescontent/actions-access-control.mdxasset29,579 bytescontent/actions-advanced.mdxasset12,329 bytescontent/actions-agent-tools.mdxasset5,783 bytescontent/actions-defining.mdxasset27,635 bytescontent/actions-other-surfaces.mdxasset10,551 bytescontent/actions-overview.mdxasset8,513 bytescontent/actions-run-context.mdxasset9,338 bytescontent/add-ui-to-your-agent.mdxasset5,098 bytescontent/administered-deployments.mdxasset7,852 bytescontent/agent-mentions.mdxasset10,425 bytescontent/agent-native-config.mdxasset33,664 bytescontent/agent-native-toolkit.mdxasset17,692 bytescontent/agent-resources.mdxasset26,264 bytescontent/agent-surfaces.mdxasset34,561 bytescontent/agent-teams.mdxasset13,735 bytescontent/agent-web-surfaces.mdxasset8,636 bytescontent/agentkit.mdxasset14,368 bytescontent/audit-log.mdxasset14,614 bytescontent/authentication.mdxasset30,722 bytescontent/automation-connectors.mdxasset8,027 bytescontent/automations.mdxasset27,305 bytescontent/aws-amplify.mdxasset10,959 bytescontent/aws-lambda.mdxasset14,193 bytescontent/aws-rds.mdxasset1,773 bytescontent/azure-postgresql.mdxasset1,812 bytescontent/azure-static-web-apps.mdxasset9,286 bytescontent/blueprint-installer.mdxasset6,429 bytescontent/cli-adapters.mdxasset5,867 bytescontent/client-advanced.mdxasset13,369 bytescontent/client-agent-chat.mdxasset17,736 bytescontent/client-data.mdxasset16,191 bytescontent/client-entry-points.mdxasset1,191 bytescontent/client-overview.mdxasset11,099 bytescontent/client-routing.mdxasset5,672 bytescontent/client-sync-internals.mdxasset5,368 bytescontent/cloneable-saas.mdxasset16,984 bytescontent/cloud-sql.mdxasset2,350 bytescontent/cloudflare.mdxasset8,477 bytescontent/code-agents-ui.mdxasset27,223 bytescontent/components.mdxasset43,057 bytescontent/context-awareness.mdxasset25,082 bytescontent/create-with-an-agent.mdxasset8,654 bytescontent/creating-templates.mdxasset22,077 bytescontent/cross-app-connect-design.mdxasset8,612 bytescontent/cross-app-sso.mdxasset8,887 bytescontent/custom-design-system.mdxasset17,756 bytescontent/data-programs.mdxasset25,607 bytescontent/decisions.mdxasset8,134 bytescontent/deno-deploy.mdxasset7,365 bytescontent/deploy-an-app.mdxasset5,076 bytescontent/deployment-environment-variables.mdxasset16,251 bytescontent/deployment.mdxasset11,536 bytescontent/dispatch.mdxasset19,002 bytescontent/docker.mdxasset8,248 bytescontent/docs-components.mdxasset29,637 bytescontent/doctor.mdxasset21,794 bytescontent/drafts/clips-experimental-features.mddoc5,575 bytescontent/drop-in-agent.mdxasset16,347 bytescontent/durable-background-runs.mdxasset14,123 bytescontent/durable-resume.mdxasset8,494 bytescontent/embedding-sdk.mdxasset31,234 bytescontent/environment-variables.mdxasset29,659 bytescontent/evals.mdxasset12,682 bytescontent/extensions.mdxasset35,802 bytescontent/external-agents-catalog.mdxasset40,861 bytescontent/external-agents.mdxasset29,074 bytescontent/file-uploads.mdxasset10,946 bytescontent/generative-ui.mdxasset11,044 bytescontent/getting-started-actions.mdxasset2,749 bytescontent/getting-started-pages.mdxasset2,655 bytescontent/getting-started.mdxasset7,189 bytescontent/harness-agents.mdxasset17,330 bytescontent/http-api.mdxasset21,430 bytescontent/integrations.mdxasset34,619 bytescontent/internationalization.mdxasset6,333 bytescontent/key-concepts.mdxasset31,343 bytescontent/koyeb.mdxasset7,497 bytescontent/locales/ar-SA/a2a-protocol.mdxasset49,622 bytescontent/locales/ar-SA/actions-access-control.mdxasset41,738 bytescontent/locales/ar-SA/actions-advanced.mdxasset14,935 bytescontent/locales/ar-SA/actions-agent-tools.mdxasset8,309 bytescontent/locales/ar-SA/actions-defining.mdxasset37,013 bytescontent/locales/ar-SA/actions-other-surfaces.mdxasset14,883 bytescontent/locales/ar-SA/actions-overview.mdxasset11,549 bytescontent/locales/ar-SA/actions-run-context.mdxasset11,412 bytescontent/locales/ar-SA/add-ui-to-your-agent.mdxasset3,775 bytescontent/locales/ar-SA/administered-deployments.mdxasset6,604 bytescontent/locales/ar-SA/agent-mentions.mdxasset12,099 bytescontent/locales/ar-SA/agent-native-toolkit.mdxasset17,719 bytescontent/locales/ar-SA/agent-resources.mdxasset35,839 bytescontent/locales/ar-SA/agent-surfaces.mdxasset49,533 bytescontent/locales/ar-SA/agent-teams.mdxasset18,994 bytescontent/locales/ar-SA/agent-web-surfaces.mdxasset11,611 bytescontent/locales/ar-SA/audit-log.mdxasset16,258 bytescontent/locales/ar-SA/authentication.mdxasset39,820 bytescontent/locales/ar-SA/automations.mdxasset32,064 bytescontent/locales/ar-SA/aws-amplify.mdxasset15,299 bytescontent/locales/ar-SA/aws-lambda.mdxasset19,327 bytescontent/locales/ar-SA/aws-rds.mdxasset319 bytescontent/locales/ar-SA/azure-postgresql.mdxasset344 bytescontent/locales/ar-SA/azure-static-web-apps.mdxasset12,736 bytescontent/locales/ar-SA/blueprint-installer.mdxasset8,563 bytescontent/locales/ar-SA/cli-adapters.mdxasset6,832 bytescontent/locales/ar-SA/client-advanced.mdxasset18,176 bytescontent/locales/ar-SA/client-agent-chat.mdxasset23,093 bytescontent/locales/ar-SA/client-data.mdxasset19,793 bytescontent/locales/ar-SA/client-entry-points.mdxasset1,633 bytescontent/locales/ar-SA/client-overview.mdxasset14,530 bytescontent/locales/ar-SA/client-routing.mdxasset7,175 bytescontent/locales/ar-SA/client-sync-internals.mdxasset7,435 bytescontent/locales/ar-SA/cloneable-saas.mdxasset23,765 bytescontent/locales/ar-SA/cloud-sql.mdxasset317 bytescontent/locales/ar-SA/cloudflare.mdxasset12,209 bytescontent/locales/ar-SA/code-agents-ui.mdxasset34,988 bytescontent/locales/ar-SA/components.mdxasset51,437 bytescontent/locales/ar-SA/context-awareness.mdxasset33,388 bytescontent/locales/ar-SA/create-with-an-agent.mdxasset12,046 bytescontent/locales/ar-SA/creating-templates.mdxasset27,329 bytescontent/locales/ar-SA/cross-app-connect-design.mdxasset1,509 bytescontent/locales/ar-SA/cross-app-sso.mdxasset9,969 bytescontent/locales/ar-SA/custom-design-system.mdxasset22,154 bytescontent/locales/ar-SA/data-programs.mdxasset34,082 bytescontent/locales/ar-SA/decisions.mdxasset9,293 bytescontent/locales/ar-SA/deno-deploy.mdxasset10,352 bytescontent/locales/ar-SA/deploy-an-app.mdxasset7,267 bytescontent/locales/ar-SA/deployment-environment-variables.mdxasset21,348 bytescontent/locales/ar-SA/deployment.mdxasset14,249 bytescontent/locales/ar-SA/dispatch.mdxasset27,771 bytescontent/locales/ar-SA/docker.mdxasset11,617 bytescontent/locales/ar-SA/doctor.mdxasset3,481 bytescontent/locales/ar-SA/drop-in-agent.mdxasset21,339 bytescontent/locales/ar-SA/durable-resume.mdxasset12,965 bytescontent/locales/ar-SA/embedding-sdk.mdxasset41,171 bytescontent/locales/ar-SA/evals.mdxasset13,665 bytescontent/locales/ar-SA/extensions.mdxasset35,021 bytescontent/locales/ar-SA/external-agents.mdxasset90,133 bytescontent/locales/ar-SA/file-uploads.mdxasset13,728 bytescontent/locales/ar-SA/getting-started-actions.mdxasset3,226 bytescontent/locales/ar-SA/getting-started-pages.mdxasset3,315 bytescontent/locales/ar-SA/getting-started.mdxasset9,642 bytescontent/locales/ar-SA/harness-agents.mdxasset22,868 bytescontent/locales/ar-SA/internationalization.mdxasset8,408 bytescontent/locales/ar-SA/key-concepts.mdxasset42,987 bytescontent/locales/ar-SA/koyeb.mdxasset10,468 bytescontent/locales/ar-SA/mcp-apps.mdxasset34,529 bytescontent/locales/ar-SA/mcp-clients.mdxasset35,855 bytescontent/locales/ar-SA/mcp-protocol.mdxasset28,300 bytescontent/locales/ar-SA/messaging.mdxasset60,655 bytescontent/locales/ar-SA/multi-app-workspace.mdxasset35,603 bytescontent/locales/ar-SA/multi-tenancy.mdxasset7,641 bytescontent/locales/ar-SA/native-chat-ui.mdxasset21,775 bytescontent/locales/ar-SA/neon.mdxasset244 bytescontent/locales/ar-SA/netlify.mdxasset14,105 bytescontent/locales/ar-SA/node-docker.mdxasset878 bytescontent/locales/ar-SA/node-js.mdxasset8,585 bytescontent/locales/ar-SA/notifications.mdxasset16,605 bytescontent/locales/ar-SA/observability.mdxasset46,236 bytescontent/locales/ar-SA/observational-memory.mdxasset11,140 bytescontent/locales/ar-SA/onboarding.mdxasset17,270 bytescontent/locales/ar-SA/organizations-teams-permissions.mdxasset5,998 bytescontent/locales/ar-SA/package-lifecycle.mdxasset3,065 bytescontent/locales/ar-SA/plan-plugin.mdxasset23,584 bytescontent/locales/ar-SA/portal.mdxasset636 bytescontent/locales/ar-SA/postgres.mdxasset487 bytescontent/locales/ar-SA/pr-visual-recap.mdxasset55,229 bytescontent/locales/ar-SA/privacy-and-data-rights.mdxasset6,363 bytescontent/locales/ar-SA/processors.mdxasset8,878 bytescontent/locales/ar-SA/progress.mdxasset10,965 bytescontent/locales/ar-SA/pure-agent-apps.mdxasset5,147 bytescontent/locales/ar-SA/real-time-collaboration.mdxasset56,951 bytescontent/locales/ar-SA/recorder.mdxasset17,105 bytescontent/locales/ar-SA/recurring-jobs.mdxasset24,397 bytescontent/locales/ar-SA/render.mdxasset11,479 bytescontent/locales/ar-SA/sandbox-adapters.mdxasset18,464 bytescontent/locales/ar-SA/security.mdxasset31,951 bytescontent/locales/ar-SA/server-database.mdxasset25,182 bytescontent/locales/ar-SA/server-middleware.mdxasset8,056 bytescontent/locales/ar-SA/server-overview.mdxasset12,472 bytescontent/locales/ar-SA/server-plugins.mdxasset9,363 bytescontent/locales/ar-SA/server-routes.mdxasset14,170 bytescontent/locales/ar-SA/sharing.mdxasset27,394 bytescontent/locales/ar-SA/skills-guide.mdxasset27,328 bytescontent/locales/ar-SA/ssr-caching.mdxasset7,593 bytescontent/locales/ar-SA/supabase.mdxasset264 bytescontent/locales/ar-SA/syncing-template-changes.mdxasset12,027 bytescontent/locales/ar-SA/template-analytics.mdxasset26,731 bytescontent/locales/ar-SA/template-assets.mdxasset32,628 bytescontent/locales/ar-SA/template-brain.mdxasset30,460 bytescontent/locales/ar-SA/template-calendar-agent.mdxasset10,416 bytescontent/locales/ar-SA/template-calendar-developers.mdxasset34,165 bytescontent/locales/ar-SA/template-calendar-features.mdxasset14,752 bytescontent/locales/ar-SA/template-calendar-integrations.mdxasset6,353 bytescontent/locales/ar-SA/template-calendar.mdxasset4,507 bytescontent/locales/ar-SA/template-chat-agent.mdxasset4,772 bytescontent/locales/ar-SA/template-chat-developers.mdxasset32,064 bytescontent/locales/ar-SA/template-chat-features.mdxasset10,112 bytescontent/locales/ar-SA/template-chat-integrations.mdxasset5,282 bytescontent/locales/ar-SA/template-chat.mdxasset5,878 bytescontent/locales/ar-SA/template-clips-agent.mdxasset6,548 bytescontent/locales/ar-SA/template-clips-developers.mdxasset36,983 bytescontent/locales/ar-SA/template-clips-embed.mdxasset14,911 bytescontent/locales/ar-SA/template-clips-features.mdxasset31,551 bytescontent/locales/ar-SA/template-clips-integrations.mdxasset13,786 bytescontent/locales/ar-SA/template-clips.mdxasset6,260 bytescontent/locales/ar-SA/template-content-local-files.mdxasset6,309 bytescontent/locales/ar-SA/template-content.mdxasset9,119 bytescontent/locales/ar-SA/template-design-agent.mdxasset8,962 bytescontent/locales/ar-SA/template-design-developers.mdxasset47,581 bytescontent/locales/ar-SA/template-design-features.mdxasset21,185 bytescontent/locales/ar-SA/template-design-integrations.mdxasset6,846 bytescontent/locales/ar-SA/template-design.mdxasset5,038 bytescontent/locales/ar-SA/template-dispatch.mdxasset14,395 bytescontent/locales/ar-SA/template-forms-agent.mdxasset5,261 bytescontent/locales/ar-SA/template-forms-features.mdxasset17,317 bytescontent/locales/ar-SA/template-forms-integrations.mdxasset6,513 bytescontent/locales/ar-SA/template-forms.mdxasset10,605 bytescontent/locales/ar-SA/template-mail-inbox.mdxasset9,437 bytescontent/locales/ar-SA/template-mail.mdxasset24,507 bytescontent/locales/ar-SA/template-plan.mdxasset15,972 bytescontent/locales/ar-SA/template-slides-agent.mdxasset5,155 bytescontent/locales/ar-SA/template-slides-developers.mdxasset27,374 bytescontent/locales/ar-SA/template-slides-features.mdxasset15,343 bytescontent/locales/ar-SA/template-slides-integrations.mdxasset5,850 bytescontent/locales/ar-SA/template-slides.mdxasset8,159 bytescontent/locales/ar-SA/toolkit-capability-packages.mdxasset1,270 bytescontent/locales/ar-SA/toolkit-context-knowledge.mdxasset1,502 bytescontent/locales/ar-SA/toolkit-editors-canvases.mdxasset3,883 bytescontent/locales/ar-SA/toolkit-settings.mdxasset28,132 bytescontent/locales/ar-SA/tracking.mdxasset28,360 bytescontent/locales/ar-SA/turn-into-app.mdxasset20,527 bytescontent/locales/ar-SA/updating-ui-in-production.mdxasset3,786 bytescontent/locales/ar-SA/upgrading-core-ui.mdxasset25,733 bytescontent/locales/ar-SA/using-your-agent.mdxasset7,784 bytescontent/locales/ar-SA/vercel.mdxasset11,728 bytescontent/locales/ar-SA/versioning-and-stability.mdxasset2,602 bytescontent/locales/ar-SA/voice-input.mdxasset10,283 bytescontent/locales/ar-SA/webmcp.mdxasset3,290 bytescontent/locales/ar-SA/what-is-agent-native.mdxasset6,295 bytescontent/locales/ar-SA/workspace-connections.mdxasset29,123 bytescontent/locales/ar-SA/workspace-deployment.mdxasset15,431 bytescontent/locales/ar-SA/workspace-management.mdxasset16,672 bytescontent/locales/ar-SA/writing-agent-instructions.mdxasset17,002 bytescontent/locales/de-DE/a2a-protocol.mdxasset42,895 bytescontent/locales/de-DE/actions-access-control.mdxasset34,055 bytescontent/locales/de-DE/actions-advanced.mdxasset13,420 bytescontent/locales/de-DE/actions-agent-tools.mdxasset6,653 bytescontent/locales/de-DE/actions-defining.mdxasset31,030 bytescontent/locales/de-DE/actions-other-surfaces.mdxasset11,812 bytescontent/locales/de-DE/actions-overview.mdxasset9,479 bytescontent/locales/de-DE/actions-run-context.mdxasset10,181 bytescontent/locales/de-DE/add-ui-to-your-agent.mdxasset3,283 bytescontent/locales/de-DE/administered-deployments.mdxasset5,520 bytescontent/locales/de-DE/agent-mentions.mdxasset10,325 bytescontent/locales/de-DE/agent-native-toolkit.mdxasset14,883 bytescontent/locales/de-DE/agent-resources.mdxasset29,156 bytescontent/locales/de-DE/agent-surfaces.mdxasset39,669 bytescontent/locales/de-DE/agent-teams.mdxasset15,901 bytescontent/locales/de-DE/agent-web-surfaces.mdxasset9,914 bytescontent/locales/de-DE/audit-log.mdxasset13,399 bytescontent/locales/de-DE/authentication.mdxasset33,091 bytescontent/locales/de-DE/automations.mdxasset29,192 bytescontent/locales/de-DE/aws-amplify.mdxasset13,084 bytescontent/locales/de-DE/aws-lambda.mdxasset16,656 bytescontent/locales/de-DE/aws-rds.mdxasset319 bytescontent/locales/de-DE/azure-postgresql.mdxasset344 bytescontent/locales/de-DE/azure-static-web-apps.mdxasset11,058 bytescontent/locales/de-DE/blueprint-installer.mdxasset7,358 bytescontent/locales/de-DE/cli-adapters.mdxasset6,330 bytescontent/locales/de-DE/client-advanced.mdxasset15,204 bytescontent/locales/de-DE/client-agent-chat.mdxasset20,148 bytescontent/locales/de-DE/client-data.mdxasset16,940 bytescontent/locales/de-DE/client-entry-points.mdxasset1,318 bytescontent/locales/de-DE/client-overview.mdxasset11,912 bytescontent/locales/de-DE/client-routing.mdxasset6,467 bytescontent/locales/de-DE/client-sync-internals.mdxasset6,075 bytescontent/locales/de-DE/cloneable-saas.mdxasset19,697 bytescontent/locales/de-DE/cloud-sql.mdxasset317 bytescontent/locales/de-DE/cloudflare.mdxasset10,181 bytescontent/locales/de-DE/code-agents-ui.mdxasset29,388 bytescontent/locales/de-DE/components.mdxasset45,787 bytescontent/locales/de-DE/context-awareness.mdxasset28,331 bytescontent/locales/de-DE/create-with-an-agent.mdxasset10,228 bytescontent/locales/de-DE/creating-templates.mdxasset23,275 bytescontent/locales/de-DE/cross-app-connect-design.mdxasset1,104 bytescontent/locales/de-DE/cross-app-sso.mdxasset7,656 bytescontent/locales/de-DE/custom-design-system.mdxasset19,060 bytescontent/locales/de-DE/data-programs.mdxasset29,620 bytescontent/locales/de-DE/decisions.mdxasset7,902 bytescontent/locales/de-DE/deno-deploy.mdxasset8,928 bytescontent/locales/de-DE/deploy-an-app.mdxasset6,275 bytescontent/locales/de-DE/deployment-environment-variables.mdxasset18,733 bytescontent/locales/de-DE/deployment.mdxasset12,742 bytescontent/locales/de-DE/dispatch.mdxasset22,566 bytescontent/locales/de-DE/docker.mdxasset9,859 bytescontent/locales/de-DE/doctor.mdxasset2,851 bytescontent/locales/de-DE/drop-in-agent.mdxasset18,780 bytescontent/locales/de-DE/durable-resume.mdxasset10,388 bytescontent/locales/de-DE/embedding-sdk.mdxasset35,283 bytescontent/locales/de-DE/evals.mdxasset12,330 bytescontent/locales/de-DE/extensions.mdxasset28,525 bytescontent/locales/de-DE/external-agents.mdxasset73,833 bytescontent/locales/de-DE/file-uploads.mdxasset12,193 bytescontent/locales/de-DE/getting-started-actions.mdxasset2,996 bytescontent/locales/de-DE/getting-started-pages.mdxasset2,980 bytescontent/locales/de-DE/getting-started.mdxasset8,302 bytescontent/locales/de-DE/harness-agents.mdxasset19,595 bytescontent/locales/de-DE/internationalization.mdxasset6,982 bytescontent/locales/de-DE/key-concepts.mdxasset35,507 bytescontent/locales/de-DE/koyeb.mdxasset9,051 bytescontent/locales/de-DE/mcp-apps.mdxasset28,071 bytescontent/locales/de-DE/mcp-clients.mdxasset30,971 bytescontent/locales/de-DE/mcp-protocol.mdxasset24,763 bytescontent/locales/de-DE/messaging.mdxasset51,380 bytescontent/locales/de-DE/multi-app-workspace.mdxasset28,927 bytescontent/locales/de-DE/multi-tenancy.mdxasset6,496 bytescontent/locales/de-DE/native-chat-ui.mdxasset18,596 bytescontent/locales/de-DE/neon.mdxasset244 bytescontent/locales/de-DE/netlify.mdxasset12,267 bytescontent/locales/de-DE/node-docker.mdxasset878 bytescontent/locales/de-DE/node-js.mdxasset7,331 bytescontent/locales/de-DE/notifications.mdxasset14,953 bytescontent/locales/de-DE/observability.mdxasset38,447 bytescontent/locales/de-DE/observational-memory.mdxasset9,503 bytescontent/locales/de-DE/onboarding.mdxasset14,738 bytescontent/locales/de-DE/organizations-teams-permissions.mdxasset4,622 bytescontent/locales/de-DE/package-lifecycle.mdxasset2,542 bytescontent/locales/de-DE/plan-plugin.mdxasset19,470 bytescontent/locales/de-DE/portal.mdxasset511 bytescontent/locales/de-DE/postgres.mdxasset420 bytescontent/locales/de-DE/pr-visual-recap.mdxasset46,003 bytescontent/locales/de-DE/privacy-and-data-rights.mdxasset4,959 bytescontent/locales/de-DE/processors.mdxasset7,499 bytescontent/locales/de-DE/progress.mdxasset9,684 bytescontent/locales/de-DE/pure-agent-apps.mdxasset4,415 bytescontent/locales/de-DE/real-time-collaboration.mdxasset48,617 bytescontent/locales/de-DE/recorder.mdxasset15,020 bytescontent/locales/de-DE/recurring-jobs.mdxasset20,940 bytescontent/locales/de-DE/render.mdxasset9,933 bytescontent/locales/de-DE/sandbox-adapters.mdxasset15,675 bytescontent/locales/de-DE/security.mdxasset26,644 bytescontent/locales/de-DE/server-database.mdxasset21,086 bytescontent/locales/de-DE/server-middleware.mdxasset6,376 bytescontent/locales/de-DE/server-overview.mdxasset10,095 bytescontent/locales/de-DE/server-plugins.mdxasset7,711 bytescontent/locales/de-DE/server-routes.mdxasset12,029 bytescontent/locales/de-DE/sharing.mdxasset23,417 bytescontent/locales/de-DE/skills-guide.mdxasset23,306 bytescontent/locales/de-DE/ssr-caching.mdxasset6,078 bytescontent/locales/de-DE/supabase.mdxasset264 bytescontent/locales/de-DE/syncing-template-changes.mdxasset9,983 bytescontent/locales/de-DE/template-analytics.mdxasset21,679 bytescontent/locales/de-DE/template-assets.mdxasset27,310 bytescontent/locales/de-DE/template-brain.mdxasset25,567 bytescontent/locales/de-DE/template-calendar-agent.mdxasset8,415 bytescontent/locales/de-DE/template-calendar-developers.mdxasset30,434 bytescontent/locales/de-DE/template-calendar-features.mdxasset11,988 bytescontent/locales/de-DE/template-calendar-integrations.mdxasset5,061 bytescontent/locales/de-DE/template-calendar.mdxasset3,694 bytescontent/locales/de-DE/template-chat-agent.mdxasset4,179 bytescontent/locales/de-DE/template-chat-developers.mdxasset28,186 bytescontent/locales/de-DE/template-chat-features.mdxasset8,332 bytescontent/locales/de-DE/template-chat-integrations.mdxasset4,107 bytescontent/locales/de-DE/template-chat.mdxasset4,816 bytescontent/locales/de-DE/template-clips-agent.mdxasset5,390 bytescontent/locales/de-DE/template-clips-developers.mdxasset32,946 bytescontent/locales/de-DE/template-clips-embed.mdxasset12,111 bytescontent/locales/de-DE/template-clips-features.mdxasset25,806 bytescontent/locales/de-DE/template-clips-integrations.mdxasset11,377 bytescontent/locales/de-DE/template-clips.mdxasset5,182 bytescontent/locales/de-DE/template-content-local-files.mdxasset6,398 bytescontent/locales/de-DE/template-content.mdxasset7,490 bytescontent/locales/de-DE/template-design-agent.mdxasset7,362 bytescontent/locales/de-DE/template-design-developers.mdxasset43,669 bytescontent/locales/de-DE/template-design-features.mdxasset17,091 bytescontent/locales/de-DE/template-design-integrations.mdxasset5,645 bytescontent/locales/de-DE/template-design.mdxasset4,094 bytescontent/locales/de-DE/template-dispatch.mdxasset11,486 bytescontent/locales/de-DE/template-forms-agent.mdxasset4,203 bytescontent/locales/de-DE/template-forms-features.mdxasset14,399 bytescontent/locales/de-DE/template-forms-integrations.mdxasset5,005 bytescontent/locales/de-DE/template-forms.mdxasset8,874 bytescontent/locales/de-DE/template-mail-inbox.mdxasset7,391 bytescontent/locales/de-DE/template-mail.mdxasset20,256 bytescontent/locales/de-DE/template-plan.mdxasset13,146 bytescontent/locales/de-DE/template-slides-agent.mdxasset4,058 bytescontent/locales/de-DE/template-slides-developers.mdxasset23,202 bytescontent/locales/de-DE/template-slides-features.mdxasset12,587 bytescontent/locales/de-DE/template-slides-integrations.mdxasset4,537 bytescontent/locales/de-DE/template-slides.mdxasset6,480 bytescontent/locales/de-DE/toolkit-capability-packages.mdxasset1,150 bytescontent/locales/de-DE/toolkit-context-knowledge.mdxasset1,115 bytescontent/locales/de-DE/toolkit-editors-canvases.mdxasset3,469 bytescontent/locales/de-DE/toolkit-settings.mdxasset24,193 bytescontent/locales/de-DE/tracking.mdxasset22,977 bytescontent/locales/de-DE/turn-into-app.mdxasset16,321 bytescontent/locales/de-DE/updating-ui-in-production.mdxasset2,929 bytescontent/locales/de-DE/upgrading-core-ui.mdxasset23,106 bytescontent/locales/de-DE/using-your-agent.mdxasset6,787 bytescontent/locales/de-DE/vercel.mdxasset9,990 bytescontent/locales/de-DE/versioning-and-stability.mdxasset2,045 bytescontent/locales/de-DE/voice-input.mdxasset8,910 bytescontent/locales/de-DE/webmcp.mdxasset2,542 bytescontent/locales/de-DE/what-is-agent-native.mdxasset4,690 bytescontent/locales/de-DE/workspace-connections.mdxasset24,382 bytescontent/locales/de-DE/workspace-deployment.mdxasset13,741 bytescontent/locales/de-DE/workspace-management.mdxasset13,698 bytescontent/locales/de-DE/writing-agent-instructions.mdxasset14,340 bytescontent/locales/es-ES/a2a-protocol.mdxasset42,832 bytescontent/locales/es-ES/actions-access-control.mdxasset33,919 bytescontent/locales/es-ES/actions-advanced.mdxasset13,869 bytescontent/locales/es-ES/actions-agent-tools.mdxasset6,895 bytescontent/locales/es-ES/actions-defining.mdxasset30,951 bytescontent/locales/es-ES/actions-other-surfaces.mdxasset11,886 bytescontent/locales/es-ES/actions-overview.mdxasset9,502 bytescontent/locales/es-ES/actions-run-context.mdxasset10,289 bytescontent/locales/es-ES/add-ui-to-your-agent.mdxasset3,049 bytescontent/locales/es-ES/administered-deployments.mdxasset5,449 bytescontent/locales/es-ES/agent-mentions.mdxasset10,188 bytescontent/locales/es-ES/agent-native-toolkit.mdxasset15,494 bytescontent/locales/es-ES/agent-resources.mdxasset29,753 bytescontent/locales/es-ES/agent-surfaces.mdxasset41,284 bytescontent/locales/es-ES/agent-teams.mdxasset15,628 bytescontent/locales/es-ES/agent-web-surfaces.mdxasset9,974 bytescontent/locales/es-ES/audit-log.mdxasset12,834 bytescontent/locales/es-ES/authentication.mdxasset34,105 bytescontent/locales/es-ES/automations.mdxasset28,845 bytescontent/locales/es-ES/aws-amplify.mdxasset13,165 bytescontent/locales/es-ES/aws-lambda.mdxasset16,472 bytescontent/locales/es-ES/aws-rds.mdxasset319 bytescontent/locales/es-ES/azure-postgresql.mdxasset344 bytescontent/locales/es-ES/azure-static-web-apps.mdxasset11,051 bytescontent/locales/es-ES/blueprint-installer.mdxasset7,076 bytescontent/locales/es-ES/cli-adapters.mdxasset6,262 bytescontent/locales/es-ES/client-advanced.mdxasset15,034 bytescontent/locales/es-ES/client-agent-chat.mdxasset19,665 bytescontent/locales/es-ES/client-data.mdxasset17,220 bytescontent/locales/es-ES/client-entry-points.mdxasset1,325 bytescontent/locales/es-ES/client-overview.mdxasset12,540 bytescontent/locales/es-ES/client-routing.mdxasset6,217 bytescontent/locales/es-ES/client-sync-internals.mdxasset6,128 bytescontent/locales/es-ES/cloneable-saas.mdxasset19,190 bytescontent/locales/es-ES/cloud-sql.mdxasset317 bytescontent/locales/es-ES/cloudflare.mdxasset10,193 bytescontent/locales/es-ES/code-agents-ui.mdxasset29,189 bytescontent/locales/es-ES/components.mdxasset45,862 bytescontent/locales/es-ES/context-awareness.mdxasset27,753 bytescontent/locales/es-ES/create-with-an-agent.mdxasset9,697 bytescontent/locales/es-ES/creating-templates.mdxasset23,472 bytescontent/locales/es-ES/cross-app-connect-design.mdxasset1,153 bytescontent/locales/es-ES/cross-app-sso.mdxasset7,788 bytescontent/locales/es-ES/custom-design-system.mdxasset19,259 bytescontent/locales/es-ES/data-programs.mdxasset28,817 bytescontent/locales/es-ES/decisions.mdxasset7,958 bytescontent/locales/es-ES/deno-deploy.mdxasset8,883 bytescontent/locales/es-ES/deploy-an-app.mdxasset6,184 bytescontent/locales/es-ES/deployment-environment-variables.mdxasset20,818 bytescontent/locales/es-ES/deployment.mdxasset12,832 bytescontent/locales/es-ES/dispatch.mdxasset22,771 bytescontent/locales/es-ES/docker.mdxasset9,922 bytescontent/locales/es-ES/doctor.mdxasset2,820 bytescontent/locales/es-ES/drop-in-agent.mdxasset18,421 bytescontent/locales/es-ES/durable-resume.mdxasset9,975 bytescontent/locales/es-ES/embedding-sdk.mdxasset34,951 bytescontent/locales/es-ES/evals.mdxasset12,221 bytescontent/locales/es-ES/extensions.mdxasset28,620 bytescontent/locales/es-ES/external-agents.mdxasset73,707 bytescontent/locales/es-ES/file-uploads.mdxasset12,043 bytescontent/locales/es-ES/getting-started-actions.mdxasset2,958 bytescontent/locales/es-ES/getting-started-pages.mdxasset2,847 bytescontent/locales/es-ES/getting-started.mdxasset8,240 bytescontent/locales/es-ES/harness-agents.mdxasset19,430 bytescontent/locales/es-ES/internationalization.mdxasset7,257 bytescontent/locales/es-ES/key-concepts.mdxasset37,239 bytescontent/locales/es-ES/koyeb.mdxasset9,046 bytescontent/locales/es-ES/mcp-apps.mdxasset28,949 bytescontent/locales/es-ES/mcp-clients.mdxasset31,628 bytescontent/locales/es-ES/mcp-protocol.mdxasset25,174 bytescontent/locales/es-ES/messaging.mdxasset50,529 bytescontent/locales/es-ES/multi-app-workspace.mdxasset29,518 bytescontent/locales/es-ES/multi-tenancy.mdxasset6,514 bytescontent/locales/es-ES/native-chat-ui.mdxasset18,815 bytescontent/locales/es-ES/neon.mdxasset244 bytescontent/locales/es-ES/netlify.mdxasset12,295 bytescontent/locales/es-ES/node-docker.mdxasset878 bytescontent/locales/es-ES/node-js.mdxasset7,317 bytescontent/locales/es-ES/notifications.mdxasset14,647 bytescontent/locales/es-ES/observability.mdxasset38,975 bytescontent/locales/es-ES/observational-memory.mdxasset9,349 bytescontent/locales/es-ES/onboarding.mdxasset14,652 bytescontent/locales/es-ES/organizations-teams-permissions.mdxasset4,792 bytescontent/locales/es-ES/package-lifecycle.mdxasset2,608 bytescontent/locales/es-ES/plan-plugin.mdxasset19,326 bytescontent/locales/es-ES/portal.mdxasset523 bytescontent/locales/es-ES/postgres.mdxasset437 bytescontent/locales/es-ES/pr-visual-recap.mdxasset46,588 bytescontent/locales/es-ES/privacy-and-data-rights.mdxasset5,015 bytescontent/locales/es-ES/processors.mdxasset7,674 bytescontent/locales/es-ES/progress.mdxasset9,589 bytescontent/locales/es-ES/pure-agent-apps.mdxasset4,401 bytescontent/locales/es-ES/real-time-collaboration.mdxasset49,565 bytescontent/locales/es-ES/recorder.mdxasset14,301 bytescontent/locales/es-ES/recurring-jobs.mdxasset20,918 bytescontent/locales/es-ES/render.mdxasset9,804 bytescontent/locales/es-ES/sandbox-adapters.mdxasset15,632 bytescontent/locales/es-ES/security.mdxasset26,609 bytescontent/locales/es-ES/server-database.mdxasset21,205 bytescontent/locales/es-ES/server-middleware.mdxasset6,531 bytescontent/locales/es-ES/server-overview.mdxasset10,485 bytescontent/locales/es-ES/server-plugins.mdxasset7,907 bytescontent/locales/es-ES/server-routes.mdxasset12,247 bytescontent/locales/es-ES/sharing.mdxasset23,289 bytescontent/locales/es-ES/skills-guide.mdxasset23,862 bytescontent/locales/es-ES/ssr-caching.mdxasset6,178 bytescontent/locales/es-ES/supabase.mdxasset264 bytescontent/locales/es-ES/syncing-template-changes.mdxasset9,581 bytescontent/locales/es-ES/template-analytics.mdxasset21,332 bytescontent/locales/es-ES/template-assets.mdxasset27,303 bytescontent/locales/es-ES/template-brain.mdxasset25,494 bytescontent/locales/es-ES/template-calendar-agent.mdxasset8,329 bytescontent/locales/es-ES/template-calendar-developers.mdxasset30,766 bytescontent/locales/es-ES/template-calendar-features.mdxasset11,826 bytescontent/locales/es-ES/template-calendar-integrations.mdxasset4,941 bytescontent/locales/es-ES/template-calendar.mdxasset3,813 bytescontent/locales/es-ES/template-chat-agent.mdxasset3,910 bytescontent/locales/es-ES/template-chat-developers.mdxasset27,798 bytescontent/locales/es-ES/template-chat-features.mdxasset7,821 bytescontent/locales/es-ES/template-chat-integrations.mdxasset3,925 bytescontent/locales/es-ES/template-chat.mdxasset4,701 bytescontent/locales/es-ES/template-clips-agent.mdxasset5,109 bytescontent/locales/es-ES/template-clips-developers.mdxasset33,381 bytescontent/locales/es-ES/template-clips-embed.mdxasset12,273 bytescontent/locales/es-ES/template-clips-features.mdxasset26,349 bytescontent/locales/es-ES/template-clips-integrations.mdxasset11,346 bytescontent/locales/es-ES/template-clips.mdxasset5,097 bytescontent/locales/es-ES/template-content-local-files.mdxasset8,136 bytescontent/locales/es-ES/template-content.mdxasset7,737 bytescontent/locales/es-ES/template-design-agent.mdxasset6,893 bytescontent/locales/es-ES/template-design-developers.mdxasset43,958 bytescontent/locales/es-ES/template-design-features.mdxasset16,566 bytescontent/locales/es-ES/template-design-integrations.mdxasset5,423 bytescontent/locales/es-ES/template-design.mdxasset3,988 bytescontent/locales/es-ES/template-dispatch.mdxasset11,845 bytescontent/locales/es-ES/template-forms-agent.mdxasset4,044 bytescontent/locales/es-ES/template-forms-features.mdxasset14,199 bytescontent/locales/es-ES/template-forms-integrations.mdxasset4,964 bytescontent/locales/es-ES/template-forms.mdxasset8,651 bytescontent/locales/es-ES/template-mail-inbox.mdxasset7,255 bytescontent/locales/es-ES/template-mail.mdxasset20,156 bytescontent/locales/es-ES/template-plan.mdxasset13,057 bytescontent/locales/es-ES/template-slides-agent.mdxasset3,884 bytescontent/locales/es-ES/template-slides-developers.mdxasset24,336 bytescontent/locales/es-ES/template-slides-features.mdxasset12,664 bytescontent/locales/es-ES/template-slides-integrations.mdxasset4,569 bytescontent/locales/es-ES/template-slides.mdxasset6,644 bytescontent/locales/es-ES/toolkit-capability-packages.mdxasset1,169 bytescontent/locales/es-ES/toolkit-context-knowledge.mdxasset1,179 bytescontent/locales/es-ES/toolkit-editors-canvases.mdxasset3,649 bytescontent/locales/es-ES/toolkit-settings.mdxasset24,842 bytescontent/locales/es-ES/tracking.mdxasset23,408 bytescontent/locales/es-ES/turn-into-app.mdxasset16,069 bytescontent/locales/es-ES/updating-ui-in-production.mdxasset3,045 bytescontent/locales/es-ES/upgrading-core-ui.mdxasset23,117 bytescontent/locales/es-ES/using-your-agent.mdxasset6,344 bytescontent/locales/es-ES/vercel.mdxasset9,977 bytescontent/locales/es-ES/versioning-and-stability.mdxasset2,139 bytescontent/locales/es-ES/voice-input.mdxasset8,591 bytescontent/locales/es-ES/webmcp.mdxasset2,642 bytescontent/locales/es-ES/what-is-agent-native.mdxasset4,680 bytescontent/locales/es-ES/workspace-connections.mdxasset24,958 bytescontent/locales/es-ES/workspace-deployment.mdxasset14,136 bytescontent/locales/es-ES/workspace-management.mdxasset14,462 bytescontent/locales/es-ES/writing-agent-instructions.mdxasset14,326 bytescontent/locales/fr-FR/a2a-protocol.mdxasset42,978 bytescontent/locales/fr-FR/actions-access-control.mdxasset35,089 bytescontent/locales/fr-FR/actions-advanced.mdxasset14,172 bytescontent/locales/fr-FR/actions-agent-tools.mdxasset6,862 bytescontent/locales/fr-FR/actions-defining.mdxasset31,815 bytescontent/locales/fr-FR/actions-other-surfaces.mdxasset12,259 bytescontent/locales/fr-FR/actions-overview.mdxasset9,786 bytescontent/locales/fr-FR/actions-run-context.mdxasset10,486 bytescontent/locales/fr-FR/add-ui-to-your-agent.mdxasset3,155 bytescontent/locales/fr-FR/administered-deployments.mdxasset5,929 bytescontent/locales/fr-FR/agent-mentions.mdxasset10,404 bytescontent/locales/fr-FR/agent-native-toolkit.mdxasset15,710 bytescontent/locales/fr-FR/agent-resources.mdxasset30,834 bytescontent/locales/fr-FR/agent-surfaces.mdxasset42,117 bytescontent/locales/fr-FR/agent-teams.mdxasset16,071 bytescontent/locales/fr-FR/agent-web-surfaces.mdxasset10,070 bytescontent/locales/fr-FR/audit-log.mdxasset13,499 bytescontent/locales/fr-FR/authentication.mdxasset34,789 bytescontent/locales/fr-FR/automations.mdxasset29,897 bytescontent/locales/fr-FR/aws-amplify.mdxasset13,366 bytescontent/locales/fr-FR/aws-lambda.mdxasset16,875 bytescontent/locales/fr-FR/aws-rds.mdxasset319 bytescontent/locales/fr-FR/azure-postgresql.mdxasset344 bytescontent/locales/fr-FR/azure-static-web-apps.mdxasset11,334 bytescontent/locales/fr-FR/blueprint-installer.mdxasset7,249 bytescontent/locales/fr-FR/cli-adapters.mdxasset6,376 bytescontent/locales/fr-FR/client-advanced.mdxasset15,418 bytescontent/locales/fr-FR/client-agent-chat.mdxasset20,274 bytescontent/locales/fr-FR/client-data.mdxasset17,596 bytescontent/locales/fr-FR/client-entry-points.mdxasset1,372 bytescontent/locales/fr-FR/client-overview.mdxasset12,635 bytescontent/locales/fr-FR/client-routing.mdxasset6,353 bytescontent/locales/fr-FR/client-sync-internals.mdxasset6,265 bytescontent/locales/fr-FR/cloneable-saas.mdxasset20,185 bytescontent/locales/fr-FR/cloud-sql.mdxasset317 bytescontent/locales/fr-FR/cloudflare.mdxasset10,494 bytescontent/locales/fr-FR/code-agents-ui.mdxasset30,173 bytescontent/locales/fr-FR/components.mdxasset47,080 bytescontent/locales/fr-FR/context-awareness.mdxasset28,715 bytescontent/locales/fr-FR/create-with-an-agent.mdxasset11,000 bytescontent/locales/fr-FR/creating-templates.mdxasset24,074 bytescontent/locales/fr-FR/cross-app-connect-design.mdxasset1,196 bytescontent/locales/fr-FR/cross-app-sso.mdxasset8,120 bytescontent/locales/fr-FR/custom-design-system.mdxasset19,947 bytescontent/locales/fr-FR/data-programs.mdxasset29,898 bytescontent/locales/fr-FR/decisions.mdxasset8,124 bytescontent/locales/fr-FR/deno-deploy.mdxasset9,155 bytescontent/locales/fr-FR/deploy-an-app.mdxasset6,470 bytescontent/locales/fr-FR/deployment-environment-variables.mdxasset19,574 bytescontent/locales/fr-FR/deployment.mdxasset13,161 bytescontent/locales/fr-FR/dispatch.mdxasset23,787 bytescontent/locales/fr-FR/docker.mdxasset10,126 bytescontent/locales/fr-FR/doctor.mdxasset3,043 bytescontent/locales/fr-FR/drop-in-agent.mdxasset19,017 bytescontent/locales/fr-FR/durable-resume.mdxasset10,187 bytescontent/locales/fr-FR/embedding-sdk.mdxasset35,666 bytescontent/locales/fr-FR/evals.mdxasset12,276 bytescontent/locales/fr-FR/extensions.mdxasset29,475 bytescontent/locales/fr-FR/external-agents.mdxasset75,876 bytescontent/locales/fr-FR/file-uploads.mdxasset12,795 bytescontent/locales/fr-FR/getting-started-actions.mdxasset3,047 bytescontent/locales/fr-FR/getting-started-pages.mdxasset2,980 bytescontent/locales/fr-FR/getting-started.mdxasset8,628 bytescontent/locales/fr-FR/harness-agents.mdxasset19,683 bytescontent/locales/fr-FR/internationalization.mdxasset7,451 bytescontent/locales/fr-FR/key-concepts.mdxasset37,828 bytescontent/locales/fr-FR/koyeb.mdxasset9,231 bytescontent/locales/fr-FR/mcp-apps.mdxasset29,784 bytescontent/locales/fr-FR/mcp-clients.mdxasset32,517 bytescontent/locales/fr-FR/mcp-protocol.mdxasset25,398 bytescontent/locales/fr-FR/messaging.mdxasset52,300 bytescontent/locales/fr-FR/multi-app-workspace.mdxasset31,111 bytescontent/locales/fr-FR/multi-tenancy.mdxasset6,658 bytescontent/locales/fr-FR/native-chat-ui.mdxasset19,287 bytescontent/locales/fr-FR/neon.mdxasset244 bytescontent/locales/fr-FR/netlify.mdxasset12,805 bytescontent/locales/fr-FR/node-docker.mdxasset878 bytescontent/locales/fr-FR/node-js.mdxasset7,509 bytescontent/locales/fr-FR/notifications.mdxasset15,028 bytescontent/locales/fr-FR/observability.mdxasset39,879 bytescontent/locales/fr-FR/observational-memory.mdxasset9,473 bytescontent/locales/fr-FR/onboarding.mdxasset15,194 bytescontent/locales/fr-FR/organizations-teams-permissions.mdxasset5,054 bytescontent/locales/fr-FR/package-lifecycle.mdxasset2,592 bytescontent/locales/fr-FR/plan-plugin.mdxasset19,895 bytescontent/locales/fr-FR/portal.mdxasset543 bytescontent/locales/fr-FR/postgres.mdxasset448 bytescontent/locales/fr-FR/pr-visual-recap.mdxasset47,923 bytescontent/locales/fr-FR/privacy-and-data-rights.mdxasset5,389 bytescontent/locales/fr-FR/processors.mdxasset7,783 bytescontent/locales/fr-FR/progress.mdxasset9,740 bytescontent/locales/fr-FR/pure-agent-apps.mdxasset4,472 bytescontent/locales/fr-FR/real-time-collaboration.mdxasset51,225 bytescontent/locales/fr-FR/recorder.mdxasset14,854 bytescontent/locales/fr-FR/recurring-jobs.mdxasset21,508 bytescontent/locales/fr-FR/render.mdxasset10,012 bytescontent/locales/fr-FR/sandbox-adapters.mdxasset16,004 bytescontent/locales/fr-FR/security.mdxasset27,440 bytescontent/locales/fr-FR/server-database.mdxasset22,001 bytescontent/locales/fr-FR/server-middleware.mdxasset6,788 bytescontent/locales/fr-FR/server-overview.mdxasset10,939 bytescontent/locales/fr-FR/server-plugins.mdxasset8,227 bytescontent/locales/fr-FR/server-routes.mdxasset12,535 bytescontent/locales/fr-FR/sharing.mdxasset24,316 bytescontent/locales/fr-FR/skills-guide.mdxasset24,533 bytescontent/locales/fr-FR/ssr-caching.mdxasset6,234 bytescontent/locales/fr-FR/supabase.mdxasset264 bytescontent/locales/fr-FR/syncing-template-changes.mdxasset10,146 bytescontent/locales/fr-FR/template-analytics.mdxasset22,530 bytescontent/locales/fr-FR/template-assets.mdxasset28,662 bytescontent/locales/fr-FR/template-brain.mdxasset26,612 bytescontent/locales/fr-FR/template-calendar-agent.mdxasset8,800 bytescontent/locales/fr-FR/template-calendar-developers.mdxasset31,647 bytescontent/locales/fr-FR/template-calendar-features.mdxasset12,610 bytescontent/locales/fr-FR/template-calendar-integrations.mdxasset5,330 bytescontent/locales/fr-FR/template-calendar.mdxasset4,061 bytescontent/locales/fr-FR/template-chat-agent.mdxasset4,269 bytescontent/locales/fr-FR/template-chat-developers.mdxasset28,633 bytescontent/locales/fr-FR/template-chat-features.mdxasset8,418 bytescontent/locales/fr-FR/template-chat-integrations.mdxasset4,261 bytescontent/locales/fr-FR/template-chat.mdxasset5,049 bytescontent/locales/fr-FR/template-clips-agent.mdxasset5,406 bytescontent/locales/fr-FR/template-clips-developers.mdxasset34,071 bytescontent/locales/fr-FR/template-clips-embed.mdxasset12,355 bytescontent/locales/fr-FR/template-clips-features.mdxasset27,581 bytescontent/locales/fr-FR/template-clips-integrations.mdxasset11,639 bytescontent/locales/fr-FR/template-clips.mdxasset5,382 bytescontent/locales/fr-FR/template-content-local-files.mdxasset6,311 bytescontent/locales/fr-FR/template-content.mdxasset7,854 bytescontent/locales/fr-FR/template-design-agent.mdxasset7,170 bytescontent/locales/fr-FR/template-design-developers.mdxasset44,751 bytescontent/locales/fr-FR/template-design-features.mdxasset17,482 bytescontent/locales/fr-FR/template-design-integrations.mdxasset5,700 bytescontent/locales/fr-FR/template-design.mdxasset4,133 bytescontent/locales/fr-FR/template-dispatch.mdxasset12,220 bytescontent/locales/fr-FR/template-forms-agent.mdxasset4,303 bytescontent/locales/fr-FR/template-forms-features.mdxasset14,938 bytescontent/locales/fr-FR/template-forms-integrations.mdxasset5,216 bytescontent/locales/fr-FR/template-forms.mdxasset9,111 bytescontent/locales/fr-FR/template-mail-inbox.mdxasset7,871 bytescontent/locales/fr-FR/template-mail.mdxasset20,727 bytescontent/locales/fr-FR/template-plan.mdxasset13,288 bytescontent/locales/fr-FR/template-slides-agent.mdxasset4,076 bytescontent/locales/fr-FR/template-slides-developers.mdxasset24,214 bytescontent/locales/fr-FR/template-slides-features.mdxasset13,226 bytescontent/locales/fr-FR/template-slides-integrations.mdxasset4,840 bytescontent/locales/fr-FR/template-slides.mdxasset6,821 bytescontent/locales/fr-FR/toolkit-capability-packages.mdxasset1,195 bytescontent/locales/fr-FR/toolkit-context-knowledge.mdxasset1,227 bytescontent/locales/fr-FR/toolkit-editors-canvases.mdxasset3,732 bytescontent/locales/fr-FR/toolkit-settings.mdxasset25,342 bytescontent/locales/fr-FR/tracking.mdxasset23,926 bytescontent/locales/fr-FR/turn-into-app.mdxasset16,864 bytescontent/locales/fr-FR/updating-ui-in-production.mdxasset3,011 bytescontent/locales/fr-FR/upgrading-core-ui.mdxasset23,693 bytescontent/locales/fr-FR/using-your-agent.mdxasset6,802 bytescontent/locales/fr-FR/vercel.mdxasset10,280 bytescontent/locales/fr-FR/versioning-and-stability.mdxasset2,330 bytescontent/locales/fr-FR/voice-input.mdxasset8,966 bytescontent/locales/fr-FR/webmcp.mdxasset2,735 bytescontent/locales/fr-FR/what-is-agent-native.mdxasset4,864 bytescontent/locales/fr-FR/workspace-connections.mdxasset26,504 bytescontent/locales/fr-FR/workspace-deployment.mdxasset13,963 bytescontent/locales/fr-FR/workspace-management.mdxasset15,099 bytescontent/locales/fr-FR/writing-agent-instructions.mdxasset14,660 bytescontent/locales/hi-IN/a2a-protocol.mdxasset64,173 bytescontent/locales/hi-IN/actions-access-control.mdxasset53,608 bytescontent/locales/hi-IN/actions-advanced.mdxasset18,984 bytescontent/locales/hi-IN/actions-agent-tools.mdxasset9,073 bytescontent/locales/hi-IN/actions-defining.mdxasset47,168 bytescontent/locales/hi-IN/actions-other-surfaces.mdxasset16,953 bytescontent/locales/hi-IN/actions-overview.mdxasset14,849 bytescontent/locales/hi-IN/actions-run-context.mdxasset13,226 bytescontent/locales/hi-IN/add-ui-to-your-agent.mdxasset3,846 bytescontent/locales/hi-IN/administered-deployments.mdxasset8,009 bytescontent/locales/hi-IN/agent-mentions.mdxasset16,248 bytescontent/locales/hi-IN/agent-native-toolkit.mdxasset21,802 bytescontent/locales/hi-IN/agent-resources.mdxasset48,218 bytescontent/locales/hi-IN/agent-surfaces.mdxasset53,127 bytescontent/locales/hi-IN/agent-teams.mdxasset24,738 bytescontent/locales/hi-IN/agent-web-surfaces.mdxasset16,116 bytescontent/locales/hi-IN/audit-log.mdxasset23,112 bytescontent/locales/hi-IN/authentication.mdxasset52,798 bytescontent/locales/hi-IN/automations.mdxasset44,057 bytescontent/locales/hi-IN/aws-amplify.mdxasset18,312 bytescontent/locales/hi-IN/aws-lambda.mdxasset22,459 bytescontent/locales/hi-IN/aws-rds.mdxasset319 bytescontent/locales/hi-IN/azure-postgresql.mdxasset344 bytescontent/locales/hi-IN/azure-static-web-apps.mdxasset15,256 bytescontent/locales/hi-IN/blueprint-installer.mdxasset12,995 bytescontent/locales/hi-IN/cli-adapters.mdxasset8,673 bytescontent/locales/hi-IN/client-advanced.mdxasset24,576 bytescontent/locales/hi-IN/client-agent-chat.mdxasset31,488 bytescontent/locales/hi-IN/client-data.mdxasset25,814 bytescontent/locales/hi-IN/client-entry-points.mdxasset2,282 bytescontent/locales/hi-IN/client-overview.mdxasset17,564 bytescontent/locales/hi-IN/client-routing.mdxasset9,484 bytescontent/locales/hi-IN/client-sync-internals.mdxasset9,636 bytescontent/locales/hi-IN/cloneable-saas.mdxasset29,685 bytescontent/locales/hi-IN/cloud-sql.mdxasset317 bytescontent/locales/hi-IN/cloudflare.mdxasset14,264 bytescontent/locales/hi-IN/code-agents-ui.mdxasset46,553 bytescontent/locales/hi-IN/components.mdxasset67,344 bytescontent/locales/hi-IN/context-awareness.mdxasset44,341 bytescontent/locales/hi-IN/create-with-an-agent.mdxasset15,158 bytescontent/locales/hi-IN/creating-templates.mdxasset36,093 bytescontent/locales/hi-IN/cross-app-connect-design.mdxasset1,388 bytescontent/locales/hi-IN/cross-app-sso.mdxasset9,101 bytescontent/locales/hi-IN/custom-design-system.mdxasset24,739 bytescontent/locales/hi-IN/data-programs.mdxasset47,419 bytescontent/locales/hi-IN/decisions.mdxasset11,948 bytescontent/locales/hi-IN/deno-deploy.mdxasset12,470 bytescontent/locales/hi-IN/deploy-an-app.mdxasset8,674 bytescontent/locales/hi-IN/deployment-environment-variables.mdxasset21,691 bytescontent/locales/hi-IN/deployment.mdxasset15,573 bytescontent/locales/hi-IN/dispatch.mdxasset40,297 bytescontent/locales/hi-IN/docker.mdxasset13,783 bytescontent/locales/hi-IN/doctor.mdxasset4,233 bytescontent/locales/hi-IN/drop-in-agent.mdxasset27,822 bytescontent/locales/hi-IN/durable-resume.mdxasset19,331 bytescontent/locales/hi-IN/embedding-sdk.mdxasset52,816 bytescontent/locales/hi-IN/evals.mdxasset17,137 bytescontent/locales/hi-IN/extensions.mdxasset49,440 bytescontent/locales/hi-IN/external-agents.mdxasset125,180 bytescontent/locales/hi-IN/file-uploads.mdxasset18,949 bytescontent/locales/hi-IN/getting-started-actions.mdxasset3,474 bytescontent/locales/hi-IN/getting-started-pages.mdxasset3,740 bytescontent/locales/hi-IN/getting-started.mdxasset10,577 bytescontent/locales/hi-IN/harness-agents.mdxasset30,971 bytescontent/locales/hi-IN/internationalization.mdxasset10,543 bytescontent/locales/hi-IN/key-concepts.mdxasset46,246 bytescontent/locales/hi-IN/koyeb.mdxasset12,935 bytescontent/locales/hi-IN/mcp-apps.mdxasset48,346 bytescontent/locales/hi-IN/mcp-clients.mdxasset51,137 bytescontent/locales/hi-IN/mcp-protocol.mdxasset37,602 bytescontent/locales/hi-IN/messaging.mdxasset83,463 bytescontent/locales/hi-IN/multi-app-workspace.mdxasset48,740 bytescontent/locales/hi-IN/multi-tenancy.mdxasset10,272 bytescontent/locales/hi-IN/native-chat-ui.mdxasset27,628 bytescontent/locales/hi-IN/neon.mdxasset244 bytescontent/locales/hi-IN/netlify.mdxasset17,124 bytescontent/locales/hi-IN/node-docker.mdxasset878 bytescontent/locales/hi-IN/node-js.mdxasset10,148 bytescontent/locales/hi-IN/notifications.mdxasset20,652 bytescontent/locales/hi-IN/observability.mdxasset62,833 bytescontent/locales/hi-IN/observational-memory.mdxasset15,766 bytescontent/locales/hi-IN/onboarding.mdxasset22,803 bytescontent/locales/hi-IN/organizations-teams-permissions.mdxasset8,414 bytescontent/locales/hi-IN/package-lifecycle.mdxasset4,004 bytescontent/locales/hi-IN/plan-plugin.mdxasset34,176 bytescontent/locales/hi-IN/portal.mdxasset758 bytescontent/locales/hi-IN/postgres.mdxasset596 bytescontent/locales/hi-IN/pr-visual-recap.mdxasset78,866 bytescontent/locales/hi-IN/privacy-and-data-rights.mdxasset6,801 bytescontent/locales/hi-IN/processors.mdxasset11,792 bytescontent/locales/hi-IN/progress.mdxasset14,511 bytescontent/locales/hi-IN/pure-agent-apps.mdxasset6,028 bytescontent/locales/hi-IN/real-time-collaboration.mdxasset79,098 bytescontent/locales/hi-IN/recorder.mdxasset24,280 bytescontent/locales/hi-IN/recurring-jobs.mdxasset31,916 bytescontent/locales/hi-IN/render.mdxasset13,917 bytescontent/locales/hi-IN/sandbox-adapters.mdxasset24,895 bytescontent/locales/hi-IN/security.mdxasset44,990 bytescontent/locales/hi-IN/server-database.mdxasset32,083 bytescontent/locales/hi-IN/server-middleware.mdxasset10,136 bytescontent/locales/hi-IN/server-overview.mdxasset15,252 bytescontent/locales/hi-IN/server-plugins.mdxasset12,616 bytescontent/locales/hi-IN/server-routes.mdxasset18,211 bytescontent/locales/hi-IN/sharing.mdxasset38,233 bytescontent/locales/hi-IN/skills-guide.mdxasset37,193 bytescontent/locales/hi-IN/ssr-caching.mdxasset8,483 bytescontent/locales/hi-IN/supabase.mdxasset264 bytescontent/locales/hi-IN/syncing-template-changes.mdxasset18,134 bytescontent/locales/hi-IN/template-analytics.mdxasset35,918 bytescontent/locales/hi-IN/template-assets.mdxasset43,188 bytescontent/locales/hi-IN/template-brain.mdxasset40,834 bytescontent/locales/hi-IN/template-calendar-agent.mdxasset15,089 bytescontent/locales/hi-IN/template-calendar-developers.mdxasset41,605 bytescontent/locales/hi-IN/template-calendar-features.mdxasset21,800 bytescontent/locales/hi-IN/template-calendar-integrations.mdxasset10,195 bytescontent/locales/hi-IN/template-calendar.mdxasset6,775 bytescontent/locales/hi-IN/template-chat-agent.mdxasset7,646 bytescontent/locales/hi-IN/template-chat-developers.mdxasset40,329 bytescontent/locales/hi-IN/template-chat-features.mdxasset15,283 bytescontent/locales/hi-IN/template-chat-integrations.mdxasset8,061 bytescontent/locales/hi-IN/template-chat.mdxasset8,776 bytescontent/locales/hi-IN/template-clips-agent.mdxasset10,562 bytescontent/locales/hi-IN/template-clips-developers.mdxasset46,033 bytescontent/locales/hi-IN/template-clips-embed.mdxasset22,937 bytescontent/locales/hi-IN/template-clips-features.mdxasset48,923 bytescontent/locales/hi-IN/template-clips-integrations.mdxasset19,814 bytescontent/locales/hi-IN/template-clips.mdxasset10,656 bytescontent/locales/hi-IN/template-content-local-files.mdxasset9,263 bytescontent/locales/hi-IN/template-content.mdxasset12,482 bytescontent/locales/hi-IN/template-design-agent.mdxasset14,165 bytescontent/locales/hi-IN/template-design-developers.mdxasset63,820 bytescontent/locales/hi-IN/template-design-features.mdxasset32,135 bytescontent/locales/hi-IN/template-design-integrations.mdxasset10,692 bytescontent/locales/hi-IN/template-design.mdxasset7,439 bytescontent/locales/hi-IN/template-dispatch.mdxasset21,596 bytescontent/locales/hi-IN/template-forms-agent.mdxasset8,581 bytescontent/locales/hi-IN/template-forms-features.mdxasset28,199 bytescontent/locales/hi-IN/template-forms-integrations.mdxasset10,435 bytescontent/locales/hi-IN/template-forms.mdxasset15,500 bytescontent/locales/hi-IN/template-mail-inbox.mdxasset12,690 bytescontent/locales/hi-IN/template-mail.mdxasset32,216 bytescontent/locales/hi-IN/template-plan.mdxasset22,885 bytescontent/locales/hi-IN/template-slides-agent.mdxasset7,776 bytescontent/locales/hi-IN/template-slides-developers.mdxasset35,107 bytescontent/locales/hi-IN/template-slides-features.mdxasset23,753 bytescontent/locales/hi-IN/template-slides-integrations.mdxasset9,169 bytescontent/locales/hi-IN/template-slides.mdxasset12,297 bytescontent/locales/hi-IN/toolkit-capability-packages.mdxasset1,275 bytescontent/locales/hi-IN/toolkit-context-knowledge.mdxasset1,424 bytescontent/locales/hi-IN/toolkit-editors-canvases.mdxasset4,263 bytescontent/locales/hi-IN/toolkit-settings.mdxasset30,201 bytescontent/locales/hi-IN/tracking.mdxasset38,476 bytescontent/locales/hi-IN/turn-into-app.mdxasset31,180 bytescontent/locales/hi-IN/updating-ui-in-production.mdxasset4,118 bytescontent/locales/hi-IN/upgrading-core-ui.mdxasset27,275 bytescontent/locales/hi-IN/using-your-agent.mdxasset11,589 bytescontent/locales/hi-IN/vercel.mdxasset14,255 bytescontent/locales/hi-IN/versioning-and-stability.mdxasset2,917 bytescontent/locales/hi-IN/voice-input.mdxasset15,365 bytescontent/locales/hi-IN/webmcp.mdxasset3,645 bytescontent/locales/hi-IN/what-is-agent-native.mdxasset7,415 bytescontent/locales/hi-IN/workspace-connections.mdxasset38,238 bytescontent/locales/hi-IN/workspace-deployment.mdxasset17,196 bytescontent/locales/hi-IN/workspace-management.mdxasset21,658 bytescontent/locales/hi-IN/writing-agent-instructions.mdxasset23,902 bytescontent/locales/ja-JP/a2a-protocol.mdxasset47,595 bytescontent/locales/ja-JP/actions-access-control.mdxasset41,261 bytescontent/locales/ja-JP/actions-advanced.mdxasset15,825 bytescontent/locales/ja-JP/actions-agent-tools.mdxasset7,783 bytescontent/locales/ja-JP/actions-defining.mdxasset37,056 bytescontent/locales/ja-JP/actions-other-surfaces.mdxasset14,140 bytescontent/locales/ja-JP/actions-overview.mdxasset11,182 bytescontent/locales/ja-JP/actions-run-context.mdxasset11,683 bytescontent/locales/ja-JP/add-ui-to-your-agent.mdxasset3,444 bytescontent/locales/ja-JP/administered-deployments.mdxasset6,039 bytescontent/locales/ja-JP/agent-mentions.mdxasset12,173 bytescontent/locales/ja-JP/agent-native-toolkit.mdxasset17,371 bytescontent/locales/ja-JP/agent-resources.mdxasset34,859 bytescontent/locales/ja-JP/agent-surfaces.mdxasset47,560 bytescontent/locales/ja-JP/agent-teams.mdxasset18,188 bytescontent/locales/ja-JP/agent-web-surfaces.mdxasset11,543 bytescontent/locales/ja-JP/audit-log.mdxasset16,002 bytescontent/locales/ja-JP/authentication.mdxasset37,641 bytescontent/locales/ja-JP/automations.mdxasset31,479 bytescontent/locales/ja-JP/aws-amplify.mdxasset15,464 bytescontent/locales/ja-JP/aws-lambda.mdxasset19,270 bytescontent/locales/ja-JP/aws-rds.mdxasset319 bytescontent/locales/ja-JP/azure-postgresql.mdxasset344 bytescontent/locales/ja-JP/azure-static-web-apps.mdxasset12,878 bytescontent/locales/ja-JP/blueprint-installer.mdxasset8,726 bytescontent/locales/ja-JP/cli-adapters.mdxasset6,900 bytescontent/locales/ja-JP/client-advanced.mdxasset17,680 bytescontent/locales/ja-JP/client-agent-chat.mdxasset22,636 bytescontent/locales/ja-JP/client-data.mdxasset19,044 bytescontent/locales/ja-JP/client-entry-points.mdxasset1,687 bytescontent/locales/ja-JP/client-overview.mdxasset13,850 bytescontent/locales/ja-JP/client-routing.mdxasset7,210 bytescontent/locales/ja-JP/client-sync-internals.mdxasset7,029 bytescontent/locales/ja-JP/cloneable-saas.mdxasset22,867 bytescontent/locales/ja-JP/cloud-sql.mdxasset317 bytescontent/locales/ja-JP/cloudflare.mdxasset11,977 bytescontent/locales/ja-JP/code-agents-ui.mdxasset34,536 bytescontent/locales/ja-JP/components.mdxasset50,793 bytescontent/locales/ja-JP/context-awareness.mdxasset32,662 bytescontent/locales/ja-JP/create-with-an-agent.mdxasset11,985 bytescontent/locales/ja-JP/creating-templates.mdxasset26,801 bytescontent/locales/ja-JP/cross-app-connect-design.mdxasset1,270 bytescontent/locales/ja-JP/cross-app-sso.mdxasset9,009 bytescontent/locales/ja-JP/custom-design-system.mdxasset21,518 bytescontent/locales/ja-JP/data-programs.mdxasset32,867 bytescontent/locales/ja-JP/decisions.mdxasset8,280 bytescontent/locales/ja-JP/deno-deploy.mdxasset10,449 bytescontent/locales/ja-JP/deploy-an-app.mdxasset7,102 bytescontent/locales/ja-JP/deployment-environment-variables.mdxasset19,937 bytescontent/locales/ja-JP/deployment.mdxasset14,478 bytescontent/locales/ja-JP/dispatch.mdxasset26,375 bytescontent/locales/ja-JP/docker.mdxasset11,797 bytescontent/locales/ja-JP/doctor.mdxasset3,297 bytescontent/locales/ja-JP/drop-in-agent.mdxasset21,319 bytescontent/locales/ja-JP/durable-resume.mdxasset11,835 bytescontent/locales/ja-JP/embedding-sdk.mdxasset39,631 bytescontent/locales/ja-JP/evals.mdxasset13,555 bytescontent/locales/ja-JP/extensions.mdxasset33,168 bytescontent/locales/ja-JP/external-agents.mdxasset86,990 bytescontent/locales/ja-JP/file-uploads.mdxasset14,085 bytescontent/locales/ja-JP/getting-started-actions.mdxasset3,249 bytescontent/locales/ja-JP/getting-started-pages.mdxasset3,387 bytescontent/locales/ja-JP/getting-started.mdxasset9,236 bytescontent/locales/ja-JP/harness-agents.mdxasset22,707 bytescontent/locales/ja-JP/internationalization.mdxasset8,176 bytescontent/locales/ja-JP/key-concepts.mdxasset41,599 bytescontent/locales/ja-JP/koyeb.mdxasset10,866 bytescontent/locales/ja-JP/mcp-apps.mdxasset33,726 bytescontent/locales/ja-JP/mcp-clients.mdxasset35,291 bytescontent/locales/ja-JP/mcp-protocol.mdxasset27,787 bytescontent/locales/ja-JP/messaging.mdxasset58,091 bytescontent/locales/ja-JP/multi-app-workspace.mdxasset33,843 bytescontent/locales/ja-JP/multi-tenancy.mdxasset7,449 bytescontent/locales/ja-JP/native-chat-ui.mdxasset21,136 bytescontent/locales/ja-JP/neon.mdxasset244 bytescontent/locales/ja-JP/netlify.mdxasset14,394 bytescontent/locales/ja-JP/node-docker.mdxasset878 bytescontent/locales/ja-JP/node-js.mdxasset8,529 bytescontent/locales/ja-JP/notifications.mdxasset15,840 bytescontent/locales/ja-JP/observability.mdxasset44,396 bytescontent/locales/ja-JP/observational-memory.mdxasset10,591 bytescontent/locales/ja-JP/onboarding.mdxasset16,543 bytescontent/locales/ja-JP/organizations-teams-permissions.mdxasset5,322 bytescontent/locales/ja-JP/package-lifecycle.mdxasset2,895 bytescontent/locales/ja-JP/plan-plugin.mdxasset23,515 bytescontent/locales/ja-JP/portal.mdxasset633 bytescontent/locales/ja-JP/postgres.mdxasset482 bytescontent/locales/ja-JP/pr-visual-recap.mdxasset55,038 bytescontent/locales/ja-JP/privacy-and-data-rights.mdxasset5,561 bytescontent/locales/ja-JP/processors.mdxasset8,337 bytescontent/locales/ja-JP/progress.mdxasset10,878 bytescontent/locales/ja-JP/pure-agent-apps.mdxasset4,859 bytescontent/locales/ja-JP/real-time-collaboration.mdxasset55,268 bytescontent/locales/ja-JP/recorder.mdxasset17,033 bytescontent/locales/ja-JP/recurring-jobs.mdxasset23,316 bytescontent/locales/ja-JP/render.mdxasset11,716 bytescontent/locales/ja-JP/sandbox-adapters.mdxasset18,536 bytescontent/locales/ja-JP/security.mdxasset30,655 bytescontent/locales/ja-JP/server-database.mdxasset24,410 bytescontent/locales/ja-JP/server-middleware.mdxasset7,525 bytescontent/locales/ja-JP/server-overview.mdxasset11,826 bytescontent/locales/ja-JP/server-plugins.mdxasset9,337 bytescontent/locales/ja-JP/server-routes.mdxasset13,913 bytescontent/locales/ja-JP/sharing.mdxasset26,374 bytescontent/locales/ja-JP/skills-guide.mdxasset27,549 bytescontent/locales/ja-JP/ssr-caching.mdxasset7,211 bytescontent/locales/ja-JP/supabase.mdxasset264 bytescontent/locales/ja-JP/syncing-template-changes.mdxasset11,485 bytescontent/locales/ja-JP/template-analytics.mdxasset24,655 bytescontent/locales/ja-JP/template-assets.mdxasset31,759 bytescontent/locales/ja-JP/template-brain.mdxasset27,950 bytescontent/locales/ja-JP/template-calendar-agent.mdxasset10,163 bytescontent/locales/ja-JP/template-calendar-developers.mdxasset33,039 bytescontent/locales/ja-JP/template-calendar-features.mdxasset13,730 bytescontent/locales/ja-JP/template-calendar-integrations.mdxasset6,354 bytescontent/locales/ja-JP/template-calendar.mdxasset4,323 bytescontent/locales/ja-JP/template-chat-agent.mdxasset5,056 bytescontent/locales/ja-JP/template-chat-developers.mdxasset32,137 bytescontent/locales/ja-JP/template-chat-features.mdxasset9,578 bytescontent/locales/ja-JP/template-chat-integrations.mdxasset5,116 bytescontent/locales/ja-JP/template-chat.mdxasset5,571 bytescontent/locales/ja-JP/template-clips-agent.mdxasset6,810 bytescontent/locales/ja-JP/template-clips-developers.mdxasset35,218 bytescontent/locales/ja-JP/template-clips-embed.mdxasset14,743 bytescontent/locales/ja-JP/template-clips-features.mdxasset29,684 bytescontent/locales/ja-JP/template-clips-integrations.mdxasset13,398 bytescontent/locales/ja-JP/template-clips.mdxasset6,360 bytescontent/locales/ja-JP/template-content-local-files.mdxasset5,495 bytescontent/locales/ja-JP/template-content.mdxasset8,765 bytescontent/locales/ja-JP/template-design-agent.mdxasset9,188 bytescontent/locales/ja-JP/template-design-developers.mdxasset48,713 bytescontent/locales/ja-JP/template-design-features.mdxasset21,118 bytescontent/locales/ja-JP/template-design-integrations.mdxasset6,975 bytescontent/locales/ja-JP/template-design.mdxasset5,144 bytescontent/locales/ja-JP/template-dispatch.mdxasset13,636 bytescontent/locales/ja-JP/template-forms-agent.mdxasset4,971 bytescontent/locales/ja-JP/template-forms-features.mdxasset16,353 bytescontent/locales/ja-JP/template-forms-integrations.mdxasset6,308 bytescontent/locales/ja-JP/template-forms.mdxasset9,833 bytescontent/locales/ja-JP/template-mail-inbox.mdxasset8,696 bytescontent/locales/ja-JP/template-mail.mdxasset23,488 bytescontent/locales/ja-JP/template-plan.mdxasset15,469 bytescontent/locales/ja-JP/template-slides-agent.mdxasset4,960 bytescontent/locales/ja-JP/template-slides-developers.mdxasset26,444 bytescontent/locales/ja-JP/template-slides-features.mdxasset14,958 bytescontent/locales/ja-JP/template-slides-integrations.mdxasset5,753 bytescontent/locales/ja-JP/template-slides.mdxasset8,056 bytescontent/locales/ja-JP/toolkit-capability-packages.mdxasset1,300 bytescontent/locales/ja-JP/toolkit-context-knowledge.mdxasset1,334 bytescontent/locales/ja-JP/toolkit-editors-canvases.mdxasset4,093 bytescontent/locales/ja-JP/toolkit-settings.mdxasset26,217 bytescontent/locales/ja-JP/tracking.mdxasset27,094 bytescontent/locales/ja-JP/turn-into-app.mdxasset20,608 bytescontent/locales/ja-JP/updating-ui-in-production.mdxasset3,545 bytescontent/locales/ja-JP/upgrading-core-ui.mdxasset23,870 bytescontent/locales/ja-JP/using-your-agent.mdxasset7,949 bytescontent/locales/ja-JP/vercel.mdxasset11,750 bytescontent/locales/ja-JP/versioning-and-stability.mdxasset2,342 bytescontent/locales/ja-JP/voice-input.mdxasset10,006 bytescontent/locales/ja-JP/webmcp.mdxasset3,072 bytescontent/locales/ja-JP/what-is-agent-native.mdxasset5,777 bytescontent/locales/ja-JP/workspace-connections.mdxasset27,559 bytescontent/locales/ja-JP/workspace-deployment.mdxasset15,695 bytescontent/locales/ja-JP/workspace-management.mdxasset15,921 bytescontent/locales/ja-JP/writing-agent-instructions.mdxasset16,849 bytescontent/locales/ko-KR/a2a-protocol.mdxasset41,727 bytescontent/locales/ko-KR/actions-access-control.mdxasset36,060 bytescontent/locales/ko-KR/actions-advanced.mdxasset13,959 bytescontent/locales/ko-KR/actions-agent-tools.mdxasset6,912 bytescontent/locales/ko-KR/actions-defining.mdxasset32,610 bytescontent/locales/ko-KR/actions-other-surfaces.mdxasset12,486 bytescontent/locales/ko-KR/actions-overview.mdxasset10,005 bytescontent/locales/ko-KR/actions-run-context.mdxasset10,419 bytescontent/locales/ko-KR/add-ui-to-your-agent.mdxasset3,213 bytescontent/locales/ko-KR/administered-deployments.mdxasset5,399 bytescontent/locales/ko-KR/agent-mentions.mdxasset10,533 bytescontent/locales/ko-KR/agent-native-toolkit.mdxasset15,192 bytescontent/locales/ko-KR/agent-resources.mdxasset29,281 bytescontent/locales/ko-KR/agent-surfaces.mdxasset42,717 bytescontent/locales/ko-KR/agent-teams.mdxasset16,000 bytescontent/locales/ko-KR/agent-web-surfaces.mdxasset9,861 bytescontent/locales/ko-KR/audit-log.mdxasset13,404 bytescontent/locales/ko-KR/authentication.mdxasset32,358 bytescontent/locales/ko-KR/automations.mdxasset27,939 bytescontent/locales/ko-KR/aws-amplify.mdxasset13,247 bytescontent/locales/ko-KR/aws-lambda.mdxasset16,615 bytescontent/locales/ko-KR/aws-rds.mdxasset319 bytescontent/locales/ko-KR/azure-postgresql.mdxasset344 bytescontent/locales/ko-KR/azure-static-web-apps.mdxasset10,909 bytescontent/locales/ko-KR/blueprint-installer.mdxasset7,392 bytescontent/locales/ko-KR/cli-adapters.mdxasset6,361 bytescontent/locales/ko-KR/client-advanced.mdxasset15,738 bytescontent/locales/ko-KR/client-agent-chat.mdxasset20,772 bytescontent/locales/ko-KR/client-data.mdxasset17,678 bytescontent/locales/ko-KR/client-entry-points.mdxasset1,474 bytescontent/locales/ko-KR/client-overview.mdxasset12,655 bytescontent/locales/ko-KR/client-routing.mdxasset6,613 bytescontent/locales/ko-KR/client-sync-internals.mdxasset6,331 bytescontent/locales/ko-KR/cloneable-saas.mdxasset20,058 bytescontent/locales/ko-KR/cloud-sql.mdxasset317 bytescontent/locales/ko-KR/cloudflare.mdxasset10,122 bytescontent/locales/ko-KR/code-agents-ui.mdxasset28,967 bytescontent/locales/ko-KR/components.mdxasset45,392 bytescontent/locales/ko-KR/context-awareness.mdxasset28,234 bytescontent/locales/ko-KR/create-with-an-agent.mdxasset9,794 bytescontent/locales/ko-KR/creating-templates.mdxasset23,513 bytescontent/locales/ko-KR/cross-app-connect-design.mdxasset1,134 bytescontent/locales/ko-KR/cross-app-sso.mdxasset7,798 bytescontent/locales/ko-KR/custom-design-system.mdxasset18,598 bytescontent/locales/ko-KR/data-programs.mdxasset29,962 bytescontent/locales/ko-KR/decisions.mdxasset7,776 bytescontent/locales/ko-KR/deno-deploy.mdxasset8,744 bytescontent/locales/ko-KR/deploy-an-app.mdxasset6,205 bytescontent/locales/ko-KR/deployment-environment-variables.mdxasset17,063 bytescontent/locales/ko-KR/deployment.mdxasset12,964 bytescontent/locales/ko-KR/dispatch.mdxasset22,504 bytescontent/locales/ko-KR/docker.mdxasset10,011 bytescontent/locales/ko-KR/doctor.mdxasset2,996 bytescontent/locales/ko-KR/drop-in-agent.mdxasset18,618 bytescontent/locales/ko-KR/durable-resume.mdxasset10,212 bytescontent/locales/ko-KR/embedding-sdk.mdxasset35,149 bytescontent/locales/ko-KR/evals.mdxasset12,131 bytescontent/locales/ko-KR/extensions.mdxasset29,101 bytescontent/locales/ko-KR/external-agents.mdxasset74,248 bytescontent/locales/ko-KR/file-uploads.mdxasset12,411 bytescontent/locales/ko-KR/getting-started-actions.mdxasset3,048 bytescontent/locales/ko-KR/getting-started-pages.mdxasset3,143 bytescontent/locales/ko-KR/getting-started.mdxasset8,378 bytescontent/locales/ko-KR/harness-agents.mdxasset19,558 bytescontent/locales/ko-KR/internationalization.mdxasset7,006 bytescontent/locales/ko-KR/key-concepts.mdxasset37,109 bytescontent/locales/ko-KR/koyeb.mdxasset9,052 bytescontent/locales/ko-KR/mcp-apps.mdxasset28,560 bytescontent/locales/ko-KR/mcp-clients.mdxasset30,372 bytescontent/locales/ko-KR/mcp-protocol.mdxasset24,601 bytescontent/locales/ko-KR/messaging.mdxasset51,349 bytescontent/locales/ko-KR/multi-app-workspace.mdxasset28,623 bytescontent/locales/ko-KR/multi-tenancy.mdxasset6,481 bytescontent/locales/ko-KR/native-chat-ui.mdxasset18,686 bytescontent/locales/ko-KR/neon.mdxasset244 bytescontent/locales/ko-KR/netlify.mdxasset12,023 bytescontent/locales/ko-KR/node-docker.mdxasset878 bytescontent/locales/ko-KR/node-js.mdxasset7,135 bytescontent/locales/ko-KR/notifications.mdxasset14,429 bytescontent/locales/ko-KR/observability.mdxasset38,741 bytescontent/locales/ko-KR/observational-memory.mdxasset9,221 bytescontent/locales/ko-KR/onboarding.mdxasset14,355 bytescontent/locales/ko-KR/organizations-teams-permissions.mdxasset4,743 bytescontent/locales/ko-KR/package-lifecycle.mdxasset2,610 bytescontent/locales/ko-KR/plan-plugin.mdxasset20,424 bytescontent/locales/ko-KR/portal.mdxasset562 bytescontent/locales/ko-KR/postgres.mdxasset420 bytescontent/locales/ko-KR/pr-visual-recap.mdxasset47,186 bytescontent/locales/ko-KR/privacy-and-data-rights.mdxasset4,905 bytescontent/locales/ko-KR/processors.mdxasset7,513 bytescontent/locales/ko-KR/progress.mdxasset9,984 bytescontent/locales/ko-KR/pure-agent-apps.mdxasset4,366 bytescontent/locales/ko-KR/real-time-collaboration.mdxasset48,680 bytescontent/locales/ko-KR/recorder.mdxasset15,137 bytescontent/locales/ko-KR/recurring-jobs.mdxasset20,795 bytescontent/locales/ko-KR/render.mdxasset9,894 bytescontent/locales/ko-KR/sandbox-adapters.mdxasset15,972 bytescontent/locales/ko-KR/security.mdxasset26,666 bytescontent/locales/ko-KR/server-database.mdxasset22,000 bytescontent/locales/ko-KR/server-middleware.mdxasset6,695 bytescontent/locales/ko-KR/server-overview.mdxasset10,714 bytescontent/locales/ko-KR/server-plugins.mdxasset8,400 bytescontent/locales/ko-KR/server-routes.mdxasset12,510 bytescontent/locales/ko-KR/sharing.mdxasset23,532 bytescontent/locales/ko-KR/skills-guide.mdxasset23,850 bytescontent/locales/ko-KR/ssr-caching.mdxasset6,398 bytescontent/locales/ko-KR/supabase.mdxasset264 bytescontent/locales/ko-KR/syncing-template-changes.mdxasset10,295 bytescontent/locales/ko-KR/template-analytics.mdxasset21,442 bytescontent/locales/ko-KR/template-assets.mdxasset27,881 bytescontent/locales/ko-KR/template-brain.mdxasset24,845 bytescontent/locales/ko-KR/template-calendar-agent.mdxasset8,777 bytescontent/locales/ko-KR/template-calendar-developers.mdxasset30,137 bytescontent/locales/ko-KR/template-calendar-features.mdxasset12,384 bytescontent/locales/ko-KR/template-calendar-integrations.mdxasset5,430 bytescontent/locales/ko-KR/template-calendar.mdxasset3,696 bytescontent/locales/ko-KR/template-chat-agent.mdxasset4,574 bytescontent/locales/ko-KR/template-chat-developers.mdxasset28,873 bytescontent/locales/ko-KR/template-chat-features.mdxasset8,229 bytescontent/locales/ko-KR/template-chat-integrations.mdxasset4,466 bytescontent/locales/ko-KR/template-chat.mdxasset4,906 bytescontent/locales/ko-KR/template-clips-agent.mdxasset5,586 bytescontent/locales/ko-KR/template-clips-developers.mdxasset32,387 bytescontent/locales/ko-KR/template-clips-embed.mdxasset12,350 bytescontent/locales/ko-KR/template-clips-features.mdxasset25,396 bytescontent/locales/ko-KR/template-clips-integrations.mdxasset11,881 bytescontent/locales/ko-KR/template-clips.mdxasset5,389 bytescontent/locales/ko-KR/template-content-local-files.mdxasset6,075 bytescontent/locales/ko-KR/template-content.mdxasset7,757 bytescontent/locales/ko-KR/template-design-agent.mdxasset7,622 bytescontent/locales/ko-KR/template-design-developers.mdxasset42,421 bytescontent/locales/ko-KR/template-design-features.mdxasset17,991 bytescontent/locales/ko-KR/template-design-integrations.mdxasset5,971 bytescontent/locales/ko-KR/template-design.mdxasset4,434 bytescontent/locales/ko-KR/template-dispatch.mdxasset11,486 bytescontent/locales/ko-KR/template-forms-agent.mdxasset4,253 bytescontent/locales/ko-KR/template-forms-features.mdxasset14,170 bytescontent/locales/ko-KR/template-forms-integrations.mdxasset5,271 bytescontent/locales/ko-KR/template-forms.mdxasset8,848 bytescontent/locales/ko-KR/template-mail-inbox.mdxasset7,481 bytescontent/locales/ko-KR/template-mail.mdxasset20,285 bytescontent/locales/ko-KR/template-plan.mdxasset13,799 bytescontent/locales/ko-KR/template-slides-agent.mdxasset4,323 bytescontent/locales/ko-KR/template-slides-developers.mdxasset24,094 bytescontent/locales/ko-KR/template-slides-features.mdxasset12,923 bytescontent/locales/ko-KR/template-slides-integrations.mdxasset4,996 bytescontent/locales/ko-KR/template-slides.mdxasset6,828 bytescontent/locales/ko-KR/toolkit-capability-packages.mdxasset1,153 bytescontent/locales/ko-KR/toolkit-context-knowledge.mdxasset1,183 bytescontent/locales/ko-KR/toolkit-editors-canvases.mdxasset3,499 bytescontent/locales/ko-KR/toolkit-settings.mdxasset23,719 bytescontent/locales/ko-KR/tracking.mdxasset23,377 bytescontent/locales/ko-KR/turn-into-app.mdxasset17,186 bytescontent/locales/ko-KR/updating-ui-in-production.mdxasset3,101 bytescontent/locales/ko-KR/upgrading-core-ui.mdxasset22,734 bytescontent/locales/ko-KR/using-your-agent.mdxasset6,832 bytescontent/locales/ko-KR/vercel.mdxasset9,922 bytescontent/locales/ko-KR/versioning-and-stability.mdxasset2,220 bytescontent/locales/ko-KR/voice-input.mdxasset8,878 bytescontent/locales/ko-KR/webmcp.mdxasset2,699 bytescontent/locales/ko-KR/what-is-agent-native.mdxasset5,269 bytescontent/locales/ko-KR/workspace-connections.mdxasset23,882 bytescontent/locales/ko-KR/workspace-deployment.mdxasset13,934 bytescontent/locales/ko-KR/workspace-management.mdxasset13,532 bytescontent/locales/ko-KR/writing-agent-instructions.mdxasset13,871 bytescontent/locales/pt-BR/a2a-protocol.mdxasset40,899 bytescontent/locales/pt-BR/actions-access-control.mdxasset32,944 bytescontent/locales/pt-BR/actions-advanced.mdxasset13,316 bytescontent/locales/pt-BR/actions-agent-tools.mdxasset6,626 bytescontent/locales/pt-BR/actions-defining.mdxasset30,112 bytescontent/locales/pt-BR/actions-other-surfaces.mdxasset11,461 bytescontent/locales/pt-BR/actions-overview.mdxasset9,250 bytescontent/locales/pt-BR/actions-run-context.mdxasset10,116 bytescontent/locales/pt-BR/add-ui-to-your-agent.mdxasset2,942 bytescontent/locales/pt-BR/administered-deployments.mdxasset5,266 bytescontent/locales/pt-BR/agent-mentions.mdxasset9,889 bytescontent/locales/pt-BR/agent-native-toolkit.mdxasset15,184 bytescontent/locales/pt-BR/agent-resources.mdxasset28,854 bytescontent/locales/pt-BR/agent-surfaces.mdxasset38,762 bytescontent/locales/pt-BR/agent-teams.mdxasset15,292 bytescontent/locales/pt-BR/agent-web-surfaces.mdxasset9,666 bytescontent/locales/pt-BR/audit-log.mdxasset12,684 bytescontent/locales/pt-BR/authentication.mdxasset32,519 bytescontent/locales/pt-BR/automations.mdxasset27,920 bytescontent/locales/pt-BR/aws-amplify.mdxasset12,546 bytescontent/locales/pt-BR/aws-lambda.mdxasset15,865 bytescontent/locales/pt-BR/aws-rds.mdxasset319 bytescontent/locales/pt-BR/azure-postgresql.mdxasset344 bytescontent/locales/pt-BR/azure-static-web-apps.mdxasset10,450 bytescontent/locales/pt-BR/blueprint-installer.mdxasset7,149 bytescontent/locales/pt-BR/cli-adapters.mdxasset6,163 bytescontent/locales/pt-BR/client-advanced.mdxasset14,833 bytescontent/locales/pt-BR/client-agent-chat.mdxasset19,432 bytescontent/locales/pt-BR/client-data.mdxasset17,035 bytescontent/locales/pt-BR/client-entry-points.mdxasset1,320 bytescontent/locales/pt-BR/client-overview.mdxasset12,277 bytescontent/locales/pt-BR/client-routing.mdxasset6,165 bytescontent/locales/pt-BR/client-sync-internals.mdxasset6,094 bytescontent/locales/pt-BR/cloneable-saas.mdxasset18,804 bytescontent/locales/pt-BR/cloud-sql.mdxasset317 bytescontent/locales/pt-BR/cloudflare.mdxasset9,645 bytescontent/locales/pt-BR/code-agents-ui.mdxasset28,553 bytescontent/locales/pt-BR/components.mdxasset44,460 bytescontent/locales/pt-BR/context-awareness.mdxasset26,834 bytescontent/locales/pt-BR/create-with-an-agent.mdxasset9,492 bytescontent/locales/pt-BR/creating-templates.mdxasset22,945 bytescontent/locales/pt-BR/cross-app-connect-design.mdxasset1,069 bytescontent/locales/pt-BR/cross-app-sso.mdxasset7,514 bytescontent/locales/pt-BR/custom-design-system.mdxasset18,948 bytescontent/locales/pt-BR/data-programs.mdxasset28,572 bytescontent/locales/pt-BR/decisions.mdxasset7,688 bytescontent/locales/pt-BR/deno-deploy.mdxasset8,347 bytescontent/locales/pt-BR/deploy-an-app.mdxasset5,883 bytescontent/locales/pt-BR/deployment-environment-variables.mdxasset19,168 bytescontent/locales/pt-BR/deployment.mdxasset12,488 bytescontent/locales/pt-BR/dispatch.mdxasset22,187 bytescontent/locales/pt-BR/docker.mdxasset9,472 bytescontent/locales/pt-BR/doctor.mdxasset2,693 bytescontent/locales/pt-BR/drop-in-agent.mdxasset18,319 bytescontent/locales/pt-BR/durable-resume.mdxasset9,768 bytescontent/locales/pt-BR/embedding-sdk.mdxasset33,772 bytescontent/locales/pt-BR/evals.mdxasset11,980 bytescontent/locales/pt-BR/extensions.mdxasset27,660 bytescontent/locales/pt-BR/external-agents.mdxasset71,267 bytescontent/locales/pt-BR/file-uploads.mdxasset11,979 bytescontent/locales/pt-BR/getting-started-actions.mdxasset2,902 bytescontent/locales/pt-BR/getting-started-pages.mdxasset2,854 bytescontent/locales/pt-BR/getting-started.mdxasset7,771 bytescontent/locales/pt-BR/harness-agents.mdxasset19,163 bytescontent/locales/pt-BR/internationalization.mdxasset6,766 bytescontent/locales/pt-BR/key-concepts.mdxasset35,570 bytescontent/locales/pt-BR/koyeb.mdxasset8,644 bytescontent/locales/pt-BR/mcp-apps.mdxasset27,630 bytescontent/locales/pt-BR/mcp-clients.mdxasset30,493 bytescontent/locales/pt-BR/mcp-protocol.mdxasset24,526 bytescontent/locales/pt-BR/messaging.mdxasset49,381 bytescontent/locales/pt-BR/multi-app-workspace.mdxasset28,624 bytescontent/locales/pt-BR/multi-tenancy.mdxasset6,344 bytescontent/locales/pt-BR/native-chat-ui.mdxasset18,497 bytescontent/locales/pt-BR/neon.mdxasset244 bytescontent/locales/pt-BR/netlify.mdxasset11,701 bytescontent/locales/pt-BR/node-docker.mdxasset878 bytescontent/locales/pt-BR/node-js.mdxasset6,980 bytescontent/locales/pt-BR/notifications.mdxasset14,278 bytescontent/locales/pt-BR/observability.mdxasset38,034 bytescontent/locales/pt-BR/observational-memory.mdxasset9,039 bytescontent/locales/pt-BR/onboarding.mdxasset14,280 bytescontent/locales/pt-BR/organizations-teams-permissions.mdxasset4,729 bytescontent/locales/pt-BR/package-lifecycle.mdxasset2,474 bytescontent/locales/pt-BR/plan-plugin.mdxasset19,053 bytescontent/locales/pt-BR/portal.mdxasset533 bytescontent/locales/pt-BR/postgres.mdxasset424 bytescontent/locales/pt-BR/pr-visual-recap.mdxasset45,490 bytescontent/locales/pt-BR/privacy-and-data-rights.mdxasset4,761 bytescontent/locales/pt-BR/processors.mdxasset7,515 bytescontent/locales/pt-BR/progress.mdxasset9,466 bytescontent/locales/pt-BR/pure-agent-apps.mdxasset4,279 bytescontent/locales/pt-BR/real-time-collaboration.mdxasset48,216 bytescontent/locales/pt-BR/recorder.mdxasset14,027 bytescontent/locales/pt-BR/recurring-jobs.mdxasset20,254 bytescontent/locales/pt-BR/render.mdxasset9,383 bytescontent/locales/pt-BR/sandbox-adapters.mdxasset15,328 bytescontent/locales/pt-BR/security.mdxasset26,125 bytescontent/locales/pt-BR/server-database.mdxasset20,978 bytescontent/locales/pt-BR/server-middleware.mdxasset6,507 bytescontent/locales/pt-BR/server-overview.mdxasset10,294 bytescontent/locales/pt-BR/server-plugins.mdxasset7,769 bytescontent/locales/pt-BR/server-routes.mdxasset12,131 bytescontent/locales/pt-BR/sharing.mdxasset23,122 bytescontent/locales/pt-BR/skills-guide.mdxasset23,122 bytescontent/locales/pt-BR/ssr-caching.mdxasset5,933 bytescontent/locales/pt-BR/supabase.mdxasset264 bytescontent/locales/pt-BR/syncing-template-changes.mdxasset9,611 bytescontent/locales/pt-BR/template-analytics.mdxasset20,988 bytescontent/locales/pt-BR/template-assets.mdxasset26,658 bytescontent/locales/pt-BR/template-brain.mdxasset25,071 bytescontent/locales/pt-BR/template-calendar-agent.mdxasset8,352 bytescontent/locales/pt-BR/template-calendar-developers.mdxasset30,149 bytescontent/locales/pt-BR/template-calendar-features.mdxasset11,905 bytescontent/locales/pt-BR/template-calendar-integrations.mdxasset4,969 bytescontent/locales/pt-BR/template-calendar.mdxasset3,812 bytescontent/locales/pt-BR/template-chat-agent.mdxasset3,885 bytescontent/locales/pt-BR/template-chat-developers.mdxasset27,550 bytescontent/locales/pt-BR/template-chat-features.mdxasset7,768 bytescontent/locales/pt-BR/template-chat-integrations.mdxasset3,954 bytescontent/locales/pt-BR/template-chat.mdxasset4,648 bytescontent/locales/pt-BR/template-clips-agent.mdxasset5,013 bytescontent/locales/pt-BR/template-clips-developers.mdxasset33,006 bytescontent/locales/pt-BR/template-clips-embed.mdxasset12,113 bytescontent/locales/pt-BR/template-clips-features.mdxasset25,730 bytescontent/locales/pt-BR/template-clips-integrations.mdxasset11,227 bytescontent/locales/pt-BR/template-clips.mdxasset5,056 bytescontent/locales/pt-BR/template-content-local-files.mdxasset6,935 bytescontent/locales/pt-BR/template-content.mdxasset7,760 bytescontent/locales/pt-BR/template-design-agent.mdxasset6,671 bytescontent/locales/pt-BR/template-design-developers.mdxasset43,057 bytescontent/locales/pt-BR/template-design-features.mdxasset16,071 bytescontent/locales/pt-BR/template-design-integrations.mdxasset5,277 bytescontent/locales/pt-BR/template-design.mdxasset3,927 bytescontent/locales/pt-BR/template-dispatch.mdxasset11,504 bytescontent/locales/pt-BR/template-forms-agent.mdxasset3,999 bytescontent/locales/pt-BR/template-forms-features.mdxasset13,998 bytescontent/locales/pt-BR/template-forms-integrations.mdxasset4,878 bytescontent/locales/pt-BR/template-forms.mdxasset8,702 bytescontent/locales/pt-BR/template-mail-inbox.mdxasset7,125 bytescontent/locales/pt-BR/template-mail.mdxasset19,588 bytescontent/locales/pt-BR/template-plan.mdxasset12,872 bytescontent/locales/pt-BR/template-slides-agent.mdxasset3,723 bytescontent/locales/pt-BR/template-slides-developers.mdxasset23,265 bytescontent/locales/pt-BR/template-slides-features.mdxasset12,020 bytescontent/locales/pt-BR/template-slides-integrations.mdxasset4,529 bytescontent/locales/pt-BR/template-slides.mdxasset6,345 bytescontent/locales/pt-BR/toolkit-capability-packages.mdxasset1,134 bytescontent/locales/pt-BR/toolkit-context-knowledge.mdxasset1,087 bytescontent/locales/pt-BR/toolkit-editors-canvases.mdxasset3,482 bytescontent/locales/pt-BR/toolkit-settings.mdxasset23,424 bytescontent/locales/pt-BR/tracking.mdxasset22,722 bytescontent/locales/pt-BR/turn-into-app.mdxasset15,701 bytescontent/locales/pt-BR/updating-ui-in-production.mdxasset2,888 bytescontent/locales/pt-BR/upgrading-core-ui.mdxasset22,756 bytescontent/locales/pt-BR/using-your-agent.mdxasset6,479 bytescontent/locales/pt-BR/vercel.mdxasset9,414 bytescontent/locales/pt-BR/versioning-and-stability.mdxasset2,061 bytescontent/locales/pt-BR/voice-input.mdxasset8,581 bytescontent/locales/pt-BR/webmcp.mdxasset2,529 bytescontent/locales/pt-BR/what-is-agent-native.mdxasset4,492 bytescontent/locales/pt-BR/workspace-connections.mdxasset24,382 bytescontent/locales/pt-BR/workspace-deployment.mdxasset13,690 bytescontent/locales/pt-BR/workspace-management.mdxasset14,314 bytescontent/locales/pt-BR/writing-agent-instructions.mdxasset13,771 bytescontent/locales/zh-CN/a2a-protocol.mdxasset35,589 bytescontent/locales/zh-CN/actions-access-control.mdxasset30,678 bytescontent/locales/zh-CN/actions-advanced.mdxasset12,025 bytescontent/locales/zh-CN/actions-agent-tools.mdxasset5,797 bytescontent/locales/zh-CN/actions-defining.mdxasset27,727 bytescontent/locales/zh-CN/actions-other-surfaces.mdxasset10,866 bytescontent/locales/zh-CN/actions-overview.mdxasset8,702 bytescontent/locales/zh-CN/actions-run-context.mdxasset9,115 bytescontent/locales/zh-CN/add-ui-to-your-agent.mdxasset2,665 bytescontent/locales/zh-CN/administered-deployments.mdxasset4,326 bytescontent/locales/zh-CN/agent-mentions.mdxasset8,929 bytescontent/locales/zh-CN/agent-native-toolkit.mdxasset13,659 bytescontent/locales/zh-CN/agent-resources.mdxasset24,432 bytescontent/locales/zh-CN/agent-surfaces.mdxasset34,546 bytescontent/locales/zh-CN/agent-teams.mdxasset13,446 bytescontent/locales/zh-CN/agent-web-surfaces.mdxasset8,557 bytescontent/locales/zh-CN/audit-log.mdxasset11,051 bytescontent/locales/zh-CN/authentication.mdxasset27,043 bytescontent/locales/zh-CN/automations.mdxasset23,799 bytescontent/locales/zh-CN/aws-amplify.mdxasset11,184 bytescontent/locales/zh-CN/aws-lambda.mdxasset14,380 bytescontent/locales/zh-CN/aws-rds.mdxasset319 bytescontent/locales/zh-CN/azure-postgresql.mdxasset344 bytescontent/locales/zh-CN/azure-static-web-apps.mdxasset9,549 bytescontent/locales/zh-CN/blueprint-installer.mdxasset6,202 bytescontent/locales/zh-CN/cli-adapters.mdxasset5,762 bytescontent/locales/zh-CN/client-advanced.mdxasset13,506 bytescontent/locales/zh-CN/client-agent-chat.mdxasset17,739 bytescontent/locales/zh-CN/client-data.mdxasset15,681 bytescontent/locales/zh-CN/client-entry-points.mdxasset1,183 bytescontent/locales/zh-CN/client-overview.mdxasset11,026 bytescontent/locales/zh-CN/client-routing.mdxasset5,606 bytescontent/locales/zh-CN/client-sync-internals.mdxasset5,553 bytescontent/locales/zh-CN/cloneable-saas.mdxasset16,797 bytescontent/locales/zh-CN/cloud-sql.mdxasset317 bytescontent/locales/zh-CN/cloudflare.mdxasset8,592 bytescontent/locales/zh-CN/code-agents-ui.mdxasset23,788 bytescontent/locales/zh-CN/components.mdxasset39,334 bytescontent/locales/zh-CN/context-awareness.mdxasset23,811 bytescontent/locales/zh-CN/create-with-an-agent.mdxasset8,538 bytescontent/locales/zh-CN/creating-templates.mdxasset20,369 bytescontent/locales/zh-CN/cross-app-connect-design.mdxasset895 bytescontent/locales/zh-CN/cross-app-sso.mdxasset6,053 bytescontent/locales/zh-CN/custom-design-system.mdxasset16,252 bytescontent/locales/zh-CN/data-programs.mdxasset24,602 bytescontent/locales/zh-CN/decisions.mdxasset6,656 bytescontent/locales/zh-CN/deno-deploy.mdxasset7,517 bytescontent/locales/zh-CN/deploy-an-app.mdxasset5,049 bytescontent/locales/zh-CN/deployment-environment-variables.mdxasset14,378 bytescontent/locales/zh-CN/deployment.mdxasset11,310 bytescontent/locales/zh-CN/dispatch.mdxasset19,030 bytescontent/locales/zh-CN/docker.mdxasset8,291 bytescontent/locales/zh-CN/doctor.mdxasset2,379 bytescontent/locales/zh-CN/drop-in-agent.mdxasset16,064 bytescontent/locales/zh-CN/durable-resume.mdxasset8,053 bytescontent/locales/zh-CN/embedding-sdk.mdxasset30,030 bytescontent/locales/zh-CN/evals.mdxasset10,752 bytescontent/locales/zh-CN/extensions.mdxasset24,055 bytescontent/locales/zh-CN/external-agents.mdxasset62,498 bytescontent/locales/zh-CN/file-uploads.mdxasset11,207 bytescontent/locales/zh-CN/getting-started-actions.mdxasset2,745 bytescontent/locales/zh-CN/getting-started-pages.mdxasset2,728 bytescontent/locales/zh-CN/getting-started.mdxasset7,022 bytescontent/locales/zh-CN/harness-agents.mdxasset16,943 bytescontent/locales/zh-CN/internationalization.mdxasset6,180 bytescontent/locales/zh-CN/key-concepts.mdxasset30,329 bytescontent/locales/zh-CN/koyeb.mdxasset7,503 bytescontent/locales/zh-CN/mcp-apps.mdxasset24,224 bytescontent/locales/zh-CN/mcp-clients.mdxasset26,673 bytescontent/locales/zh-CN/mcp-protocol.mdxasset21,492 bytescontent/locales/zh-CN/messaging.mdxasset44,478 bytescontent/locales/zh-CN/multi-app-workspace.mdxasset24,291 bytescontent/locales/zh-CN/multi-tenancy.mdxasset5,580 bytescontent/locales/zh-CN/native-chat-ui.mdxasset16,469 bytescontent/locales/zh-CN/neon.mdxasset244 bytescontent/locales/zh-CN/netlify.mdxasset9,927 bytescontent/locales/zh-CN/node-docker.mdxasset878 bytescontent/locales/zh-CN/node-js.mdxasset6,056 bytescontent/locales/zh-CN/notifications.mdxasset13,084 bytescontent/locales/zh-CN/observability.mdxasset33,480 bytescontent/locales/zh-CN/observational-memory.mdxasset7,629 bytescontent/locales/zh-CN/onboarding.mdxasset12,514 bytescontent/locales/zh-CN/organizations-teams-permissions.mdxasset3,638 bytescontent/locales/zh-CN/package-lifecycle.mdxasset2,131 bytescontent/locales/zh-CN/plan-plugin.mdxasset17,600 bytescontent/locales/zh-CN/portal.mdxasset455 bytescontent/locales/zh-CN/postgres.mdxasset371 bytescontent/locales/zh-CN/pr-visual-recap.mdxasset41,235 bytescontent/locales/zh-CN/privacy-and-data-rights.mdxasset4,004 bytescontent/locales/zh-CN/processors.mdxasset6,325 bytescontent/locales/zh-CN/progress.mdxasset8,719 bytescontent/locales/zh-CN/pure-agent-apps.mdxasset3,866 bytescontent/locales/zh-CN/real-time-collaboration.mdxasset40,268 bytescontent/locales/zh-CN/recorder.mdxasset13,123 bytescontent/locales/zh-CN/recurring-jobs.mdxasset17,780 bytescontent/locales/zh-CN/render.mdxasset8,297 bytescontent/locales/zh-CN/sandbox-adapters.mdxasset13,278 bytescontent/locales/zh-CN/security.mdxasset21,963 bytescontent/locales/zh-CN/server-database.mdxasset19,013 bytescontent/locales/zh-CN/server-middleware.mdxasset5,981 bytescontent/locales/zh-CN/server-overview.mdxasset9,327 bytescontent/locales/zh-CN/server-plugins.mdxasset6,962 bytescontent/locales/zh-CN/server-routes.mdxasset10,734 bytescontent/locales/zh-CN/sharing.mdxasset20,037 bytescontent/locales/zh-CN/skills-guide.mdxasset20,135 bytescontent/locales/zh-CN/ssr-caching.mdxasset5,500 bytescontent/locales/zh-CN/supabase.mdxasset264 bytescontent/locales/zh-CN/syncing-template-changes.mdxasset8,420 bytescontent/locales/zh-CN/template-analytics.mdxasset18,490 bytescontent/locales/zh-CN/template-assets.mdxasset23,298 bytescontent/locales/zh-CN/template-brain.mdxasset21,816 bytescontent/locales/zh-CN/template-calendar-agent.mdxasset7,768 bytescontent/locales/zh-CN/template-calendar-developers.mdxasset27,609 bytescontent/locales/zh-CN/template-calendar-features.mdxasset10,671 bytescontent/locales/zh-CN/template-calendar-integrations.mdxasset4,435 bytescontent/locales/zh-CN/template-calendar.mdxasset3,207 bytescontent/locales/zh-CN/template-chat-agent.mdxasset3,651 bytescontent/locales/zh-CN/template-chat-developers.mdxasset26,171 bytescontent/locales/zh-CN/template-chat-features.mdxasset6,705 bytescontent/locales/zh-CN/template-chat-integrations.mdxasset3,590 bytescontent/locales/zh-CN/template-chat.mdxasset4,223 bytescontent/locales/zh-CN/template-clips-agent.mdxasset4,762 bytescontent/locales/zh-CN/template-clips-developers.mdxasset29,662 bytescontent/locales/zh-CN/template-clips-embed.mdxasset10,410 bytescontent/locales/zh-CN/template-clips-features.mdxasset21,603 bytescontent/locales/zh-CN/template-clips-integrations.mdxasset10,126 bytescontent/locales/zh-CN/template-clips.mdxasset4,557 bytescontent/locales/zh-CN/template-content-local-files.mdxasset5,460 bytescontent/locales/zh-CN/template-content.mdxasset6,900 bytescontent/locales/zh-CN/template-design-agent.mdxasset6,120 bytescontent/locales/zh-CN/template-design-developers.mdxasset37,977 bytescontent/locales/zh-CN/template-design-features.mdxasset14,667 bytescontent/locales/zh-CN/template-design-integrations.mdxasset4,967 bytescontent/locales/zh-CN/template-design.mdxasset3,536 bytescontent/locales/zh-CN/template-dispatch.mdxasset10,030 bytescontent/locales/zh-CN/template-forms-agent.mdxasset3,644 bytescontent/locales/zh-CN/template-forms-features.mdxasset12,178 bytescontent/locales/zh-CN/template-forms-integrations.mdxasset4,397 bytescontent/locales/zh-CN/template-forms.mdxasset7,491 bytescontent/locales/zh-CN/template-mail-inbox.mdxasset6,609 bytescontent/locales/zh-CN/template-mail.mdxasset17,860 bytescontent/locales/zh-CN/template-plan.mdxasset11,904 bytescontent/locales/zh-CN/template-slides-agent.mdxasset3,726 bytescontent/locales/zh-CN/template-slides-developers.mdxasset22,113 bytescontent/locales/zh-CN/template-slides-features.mdxasset10,766 bytescontent/locales/zh-CN/template-slides-integrations.mdxasset4,129 bytescontent/locales/zh-CN/template-slides.mdxasset5,751 bytescontent/locales/zh-CN/toolkit-capability-packages.mdxasset961 bytescontent/locales/zh-CN/toolkit-context-knowledge.mdxasset935 bytescontent/locales/zh-CN/toolkit-editors-canvases.mdxasset3,283 bytescontent/locales/zh-CN/toolkit-settings.mdxasset21,464 bytescontent/locales/zh-CN/tracking.mdxasset20,226 bytescontent/locales/zh-CN/turn-into-app.mdxasset14,256 bytescontent/locales/zh-CN/updating-ui-in-production.mdxasset2,449 bytescontent/locales/zh-CN/upgrading-core-ui.mdxasset21,237 bytescontent/locales/zh-CN/using-your-agent.mdxasset5,949 bytescontent/locales/zh-CN/vercel.mdxasset8,240 bytescontent/locales/zh-CN/versioning-and-stability.mdxasset1,753 bytescontent/locales/zh-CN/voice-input.mdxasset7,610 bytescontent/locales/zh-CN/webmcp.mdxasset2,106 bytescontent/locales/zh-CN/what-is-agent-native.mdxasset3,962 bytescontent/locales/zh-CN/workspace-connections.mdxasset21,064 bytescontent/locales/zh-CN/workspace-deployment.mdxasset12,618 bytescontent/locales/zh-CN/workspace-management.mdxasset11,655 bytescontent/locales/zh-CN/writing-agent-instructions.mdxasset11,474 bytescontent/locales/zh-TW/a2a-protocol.mdxasset35,788 bytescontent/locales/zh-TW/actions-access-control.mdxasset31,940 bytescontent/locales/zh-TW/actions-advanced.mdxasset12,052 bytescontent/locales/zh-TW/actions-agent-tools.mdxasset6,070 bytescontent/locales/zh-TW/actions-defining.mdxasset27,977 bytescontent/locales/zh-TW/actions-other-surfaces.mdxasset11,079 bytescontent/locales/zh-TW/actions-overview.mdxasset8,881 bytescontent/locales/zh-TW/actions-run-context.mdxasset9,192 bytescontent/locales/zh-TW/add-ui-to-your-agent.mdxasset2,738 bytescontent/locales/zh-TW/administered-deployments.mdxasset4,398 bytescontent/locales/zh-TW/agent-mentions.mdxasset8,866 bytescontent/locales/zh-TW/agent-native-toolkit.mdxasset13,988 bytescontent/locales/zh-TW/agent-resources.mdxasset24,408 bytescontent/locales/zh-TW/agent-surfaces.mdxasset35,115 bytescontent/locales/zh-TW/agent-teams.mdxasset13,452 bytescontent/locales/zh-TW/agent-web-surfaces.mdxasset8,564 bytescontent/locales/zh-TW/audit-log.mdxasset11,150 bytescontent/locales/zh-TW/authentication.mdxasset27,334 bytescontent/locales/zh-TW/automations.mdxasset23,901 bytescontent/locales/zh-TW/aws-amplify.mdxasset11,645 bytescontent/locales/zh-TW/aws-lambda.mdxasset14,794 bytescontent/locales/zh-TW/aws-rds.mdxasset319 bytescontent/locales/zh-TW/azure-postgresql.mdxasset344 bytescontent/locales/zh-TW/azure-static-web-apps.mdxasset9,999 bytescontent/locales/zh-TW/blueprint-installer.mdxasset6,206 bytescontent/locales/zh-TW/cli-adapters.mdxasset5,765 bytescontent/locales/zh-TW/client-advanced.mdxasset13,335 bytescontent/locales/zh-TW/client-agent-chat.mdxasset17,675 bytescontent/locales/zh-TW/client-data.mdxasset15,635 bytescontent/locales/zh-TW/client-entry-points.mdxasset1,174 bytescontent/locales/zh-TW/client-overview.mdxasset11,135 bytescontent/locales/zh-TW/client-routing.mdxasset5,713 bytescontent/locales/zh-TW/client-sync-internals.mdxasset5,460 bytescontent/locales/zh-TW/cloneable-saas.mdxasset17,203 bytescontent/locales/zh-TW/cloud-sql.mdxasset317 bytescontent/locales/zh-TW/cloudflare.mdxasset8,871 bytescontent/locales/zh-TW/code-agents-ui.mdxasset24,139 bytescontent/locales/zh-TW/components.mdxasset39,525 bytescontent/locales/zh-TW/context-awareness.mdxasset23,887 bytescontent/locales/zh-TW/create-with-an-agent.mdxasset8,831 bytescontent/locales/zh-TW/creating-templates.mdxasset20,491 bytescontent/locales/zh-TW/cross-app-connect-design.mdxasset913 bytescontent/locales/zh-TW/cross-app-sso.mdxasset6,372 bytescontent/locales/zh-TW/custom-design-system.mdxasset15,943 bytescontent/locales/zh-TW/data-programs.mdxasset24,758 bytescontent/locales/zh-TW/decisions.mdxasset6,690 bytescontent/locales/zh-TW/deno-deploy.mdxasset7,893 bytescontent/locales/zh-TW/deploy-an-app.mdxasset5,297 bytescontent/locales/zh-TW/deployment-environment-variables.mdxasset14,564 bytescontent/locales/zh-TW/deployment.mdxasset11,494 bytescontent/locales/zh-TW/dispatch.mdxasset19,093 bytescontent/locales/zh-TW/docker.mdxasset8,711 bytescontent/locales/zh-TW/doctor.mdxasset2,400 bytescontent/locales/zh-TW/drop-in-agent.mdxasset16,055 bytescontent/locales/zh-TW/durable-background-runs.mdxasset13,634 bytescontent/locales/zh-TW/durable-resume.mdxasset8,064 bytescontent/locales/zh-TW/embedding-sdk.mdxasset30,308 bytescontent/locales/zh-TW/evals.mdxasset10,763 bytescontent/locales/zh-TW/extensions.mdxasset24,664 bytescontent/locales/zh-TW/external-agents.mdxasset62,926 bytescontent/locales/zh-TW/file-uploads.mdxasset11,308 bytescontent/locales/zh-TW/getting-started-actions.mdxasset2,719 bytescontent/locales/zh-TW/getting-started-pages.mdxasset2,716 bytescontent/locales/zh-TW/getting-started.mdxasset7,009 bytescontent/locales/zh-TW/harness-agents.mdxasset17,151 bytescontent/locales/zh-TW/internationalization.mdxasset6,252 bytescontent/locales/zh-TW/key-concepts.mdxasset31,240 bytescontent/locales/zh-TW/koyeb.mdxasset7,866 bytescontent/locales/zh-TW/mcp-apps.mdxasset24,481 bytescontent/locales/zh-TW/mcp-clients.mdxasset26,781 bytescontent/locales/zh-TW/mcp-protocol.mdxasset21,500 bytescontent/locales/zh-TW/messaging.mdxasset44,595 bytescontent/locales/zh-TW/multi-app-workspace.mdxasset24,503 bytescontent/locales/zh-TW/multi-tenancy.mdxasset5,647 bytescontent/locales/zh-TW/native-chat-ui.mdxasset16,473 bytescontent/locales/zh-TW/neon.mdxasset244 bytescontent/locales/zh-TW/netlify.mdxasset10,296 bytescontent/locales/zh-TW/node-docker.mdxasset878 bytescontent/locales/zh-TW/node-js.mdxasset6,356 bytescontent/locales/zh-TW/notifications.mdxasset13,124 bytescontent/locales/zh-TW/observability.mdxasset33,721 bytescontent/locales/zh-TW/observational-memory.mdxasset7,657 bytescontent/locales/zh-TW/onboarding.mdxasset12,513 bytescontent/locales/zh-TW/organizations-teams-permissions.mdxasset3,767 bytescontent/locales/zh-TW/package-lifecycle.mdxasset2,143 bytescontent/locales/zh-TW/plan-plugin.mdxasset17,398 bytescontent/locales/zh-TW/portal.mdxasset461 bytescontent/locales/zh-TW/postgres.mdxasset365 bytescontent/locales/zh-TW/pr-visual-recap.mdxasset41,315 bytescontent/locales/zh-TW/privacy-and-data-rights.mdxasset4,022 bytescontent/locales/zh-TW/processors.mdxasset6,340 bytescontent/locales/zh-TW/progress.mdxasset8,730 bytescontent/locales/zh-TW/pure-agent-apps.mdxasset3,899 bytescontent/locales/zh-TW/real-time-collaboration.mdxasset40,744 bytescontent/locales/zh-TW/recorder.mdxasset13,406 bytescontent/locales/zh-TW/recurring-jobs.mdxasset17,896 bytescontent/locales/zh-TW/render.mdxasset8,660 bytescontent/locales/zh-TW/sandbox-adapters.mdxasset13,343 bytescontent/locales/zh-TW/security.mdxasset22,111 bytescontent/locales/zh-TW/server-database.mdxasset19,108 bytescontent/locales/zh-TW/server-middleware.mdxasset5,872 bytescontent/locales/zh-TW/server-overview.mdxasset9,264 bytescontent/locales/zh-TW/server-plugins.mdxasset6,960 bytescontent/locales/zh-TW/server-routes.mdxasset10,860 bytescontent/locales/zh-TW/sharing.mdxasset20,150 bytescontent/locales/zh-TW/skills-guide.mdxasset20,208 bytescontent/locales/zh-TW/ssr-caching.mdxasset5,686 bytescontent/locales/zh-TW/supabase.mdxasset264 bytescontent/locales/zh-TW/syncing-template-changes.mdxasset8,590 bytescontent/locales/zh-TW/template-analytics.mdxasset18,601 bytescontent/locales/zh-TW/template-assets.mdxasset23,390 bytescontent/locales/zh-TW/template-brain.mdxasset21,890 bytescontent/locales/zh-TW/template-calendar-agent.mdxasset7,657 bytescontent/locales/zh-TW/template-calendar-developers.mdxasset27,964 bytescontent/locales/zh-TW/template-calendar-features.mdxasset10,546 bytescontent/locales/zh-TW/template-calendar-integrations.mdxasset4,396 bytescontent/locales/zh-TW/template-calendar.mdxasset3,281 bytescontent/locales/zh-TW/template-chat-agent.mdxasset3,680 bytescontent/locales/zh-TW/template-chat-developers.mdxasset26,452 bytescontent/locales/zh-TW/template-chat-features.mdxasset6,880 bytescontent/locales/zh-TW/template-chat-integrations.mdxasset3,640 bytescontent/locales/zh-TW/template-chat.mdxasset4,293 bytescontent/locales/zh-TW/template-clips-agent.mdxasset4,621 bytescontent/locales/zh-TW/template-clips-developers.mdxasset29,709 bytescontent/locales/zh-TW/template-clips-embed.mdxasset10,411 bytescontent/locales/zh-TW/template-clips-features.mdxasset21,039 bytescontent/locales/zh-TW/template-clips-integrations.mdxasset10,045 bytescontent/locales/zh-TW/template-clips.mdxasset4,579 bytescontent/locales/zh-TW/template-content-local-files.mdxasset5,608 bytescontent/locales/zh-TW/template-content.mdxasset6,933 bytescontent/locales/zh-TW/template-design-agent.mdxasset6,345 bytescontent/locales/zh-TW/template-design-developers.mdxasset39,111 bytescontent/locales/zh-TW/template-design-features.mdxasset14,830 bytescontent/locales/zh-TW/template-design-integrations.mdxasset4,895 bytescontent/locales/zh-TW/template-design.mdxasset3,562 bytescontent/locales/zh-TW/template-dispatch.mdxasset10,411 bytescontent/locales/zh-TW/template-forms-agent.mdxasset3,630 bytescontent/locales/zh-TW/template-forms-features.mdxasset12,133 bytescontent/locales/zh-TW/template-forms-integrations.mdxasset4,399 bytescontent/locales/zh-TW/template-forms.mdxasset7,587 bytescontent/locales/zh-TW/template-mail-inbox.mdxasset6,517 bytescontent/locales/zh-TW/template-mail.mdxasset17,963 bytescontent/locales/zh-TW/template-plan.mdxasset11,785 bytescontent/locales/zh-TW/template-slides-agent.mdxasset3,594 bytescontent/locales/zh-TW/template-slides-developers.mdxasset21,677 bytescontent/locales/zh-TW/template-slides-features.mdxasset10,395 bytescontent/locales/zh-TW/template-slides-integrations.mdxasset4,039 bytescontent/locales/zh-TW/template-slides.mdxasset5,642 bytescontent/locales/zh-TW/toolkit-capability-packages.mdxasset991 bytescontent/locales/zh-TW/toolkit-context-knowledge.mdxasset926 bytescontent/locales/zh-TW/toolkit-editors-canvases.mdxasset3,259 bytescontent/locales/zh-TW/toolkit-settings.mdxasset21,889 bytescontent/locales/zh-TW/tracking.mdxasset20,414 bytescontent/locales/zh-TW/turn-into-app.mdxasset14,597 bytescontent/locales/zh-TW/updating-ui-in-production.mdxasset2,559 bytescontent/locales/zh-TW/upgrading-core-ui.mdxasset21,277 bytescontent/locales/zh-TW/using-your-agent.mdxasset5,943 bytescontent/locales/zh-TW/vercel.mdxasset8,568 bytescontent/locales/zh-TW/versioning-and-stability.mdxasset1,741 bytescontent/locales/zh-TW/voice-input.mdxasset7,664 bytescontent/locales/zh-TW/webmcp.mdxasset2,103 bytescontent/locales/zh-TW/what-is-agent-native.mdxasset3,997 bytescontent/locales/zh-TW/workspace-connections.mdxasset21,129 bytescontent/locales/zh-TW/workspace-deployment.mdxasset12,956 bytescontent/locales/zh-TW/workspace-management.mdxasset11,675 bytescontent/locales/zh-TW/writing-agent-instructions.mdxasset11,440 bytescontent/mcp-apps.mdxasset24,758 bytescontent/mcp-clients.mdxasset31,294 bytescontent/mcp-protocol.mdxasset23,002 bytescontent/messaging-internals.mdxasset19,185 bytescontent/messaging-recipes.mdxasset11,255 bytescontent/messaging.mdxasset24,353 bytescontent/multi-app-workspace.mdxasset22,811 bytescontent/multi-tenancy.mdxasset5,775 bytescontent/native-chat-ui.mdxasset17,430 bytescontent/neon.mdxasset1,639 bytescontent/netlify.mdxasset10,146 bytescontent/node-docker.mdxasset851 bytescontent/node-js.mdxasset6,043 bytescontent/notifications.mdxasset13,839 bytescontent/observability.mdxasset37,581 bytescontent/observational-memory.mdxasset8,066 bytescontent/onboarding.mdxasset15,181 bytescontent/organizations-teams-permissions.mdxasset34,745 bytescontent/package-lifecycle.mdxasset6,544 bytescontent/plan-plugin.mdxasset16,715 bytescontent/portal.mdxasset4,240 bytescontent/postgres.mdxasset1,769 bytescontent/pr-visual-recap.mdxasset39,795 bytescontent/privacy-and-data-rights.mdxasset4,875 bytescontent/processors.mdxasset6,716 bytescontent/progress.mdxasset9,993 bytescontent/pure-agent-apps.mdxasset6,424 bytescontent/real-time-collaboration.mdxasset41,105 bytescontent/recorder.mdxasset12,805 bytescontent/recurring-jobs.mdxasset18,411 bytescontent/render.mdxasset8,255 bytescontent/sandbox-adapters.mdxasset13,104 bytescontent/security.mdxasset23,160 bytescontent/server-database.mdxasset19,883 bytescontent/server-middleware.mdxasset5,871 bytescontent/server-overview.mdxasset9,452 bytescontent/server-plugins.mdxasset8,223 bytescontent/server-routes.mdxasset10,903 bytescontent/sharing.mdxasset23,449 bytescontent/skills-guide.mdxasset24,577 bytescontent/ssr-caching.mdxasset5,315 bytescontent/supabase.mdxasset1,682 bytescontent/syncing-template-changes.mdxasset8,753 bytescontent/template-analytics-connectors.mdxasset7,550 bytescontent/template-analytics-dashboards.mdxasset6,843 bytescontent/template-analytics-developers.mdxasset14,487 bytescontent/template-analytics-monitoring-and-sessions.mdxasset8,520 bytescontent/template-analytics.mdxasset10,477 bytescontent/template-assets-developers.mdxasset18,635 bytescontent/template-assets-generation.mdxasset8,047 bytescontent/template-assets-integrations.mdxasset14,859 bytescontent/template-assets-presets.mdxasset10,379 bytescontent/template-assets.mdxasset9,007 bytescontent/template-brain-agent.mdxasset4,894 bytescontent/template-brain-developers.mdxasset18,239 bytescontent/template-brain-knowledge.mdxasset5,414 bytescontent/template-brain-sources.mdxasset9,567 bytescontent/template-brain.mdxasset9,564 bytescontent/template-calendar-agent.mdxasset7,248 bytescontent/template-calendar-developers.mdxasset27,444 bytescontent/template-calendar-features.mdxasset10,037 bytescontent/template-calendar-integrations.mdxasset4,195 bytescontent/template-calendar.mdxasset3,183 bytescontent/template-chat-agent.mdxasset3,472 bytescontent/template-chat-developers.mdxasset25,451 bytescontent/template-chat-features.mdxasset6,727 bytescontent/template-chat-integrations.mdxasset3,417 bytescontent/template-chat.mdxasset4,077 bytescontent/template-clips-agent.mdxasset4,445 bytescontent/template-clips-developers.mdxasset29,863 bytescontent/template-clips-embed.mdxasset10,500 bytescontent/template-clips-features.mdxasset21,114 bytescontent/template-clips-integrations.mdxasset10,003 bytescontent/template-clips.mdxasset4,308 bytescontent/template-content-databases.mdxasset15,711 bytescontent/template-content-developers.mdxasset28,275 bytescontent/template-content-editing.mdxasset5,435 bytescontent/template-content-local-files.mdxasset7,162 bytescontent/template-content-sync.mdxasset8,127 bytescontent/template-content.mdxasset6,717 bytescontent/template-design-agent.mdxasset5,873 bytescontent/template-design-developers.mdxasset38,314 bytescontent/template-design-features.mdxasset13,910 bytescontent/template-design-integrations.mdxasset4,668 bytescontent/template-design.mdxasset3,530 bytescontent/template-dispatch-agent.mdxasset7,120 bytescontent/template-dispatch-developers.mdxasset26,373 bytescontent/template-dispatch-features.mdxasset15,948 bytescontent/template-dispatch-integrations.mdxasset5,216 bytescontent/template-dispatch-reference.mdxasset23,161 bytescontent/template-dispatch.mdxasset9,426 bytescontent/template-forms-agent.mdxasset3,438 bytescontent/template-forms-developers.mdxasset25,484 bytescontent/template-forms-features.mdxasset12,048 bytescontent/template-forms-integrations.mdxasset4,145 bytescontent/template-forms.mdxasset7,507 bytescontent/template-mail-agent.mdxasset5,587 bytescontent/template-mail-developers.mdxasset18,499 bytescontent/template-mail-drafts-and-queue.mdxasset6,978 bytescontent/template-mail-inbox.mdxasset6,139 bytescontent/template-mail.mdxasset7,174 bytescontent/template-plan-automations.mdxasset9,554 bytescontent/template-plan-developers.mdxasset24,908 bytescontent/template-plan-local-and-desktop.mdxasset8,854 bytescontent/template-plan-review-workflow.mdxasset9,193 bytescontent/template-plan.mdxasset11,599 bytescontent/template-slides-agent.mdxasset3,185 bytescontent/template-slides-developers.mdxasset20,764 bytescontent/template-slides-features.mdxasset10,041 bytescontent/template-slides-integrations.mdxasset3,761 bytescontent/template-slides.mdxasset5,355 bytescontent/toolkit-agent-ux.mdxasset7,428 bytescontent/toolkit-capability-packages.mdxasset6,137 bytescontent/toolkit-collaboration.mdxasset8,100 bytescontent/toolkit-command-navigation.mdxasset6,711 bytescontent/toolkit-comments-review.mdxasset8,222 bytescontent/toolkit-context-knowledge.mdxasset5,818 bytescontent/toolkit-editors-canvases.mdxasset11,913 bytescontent/toolkit-history.mdxasset6,486 bytescontent/toolkit-observability.mdxasset5,133 bytescontent/toolkit-org-team.mdxasset10,027 bytescontent/toolkit-resources.mdxasset6,465 bytescontent/toolkit-settings.mdxasset21,545 bytescontent/toolkit-setup-connections.mdxasset5,808 bytescontent/toolkit-sharing.mdxasset13,907 bytescontent/toolkit-ui.mdxasset15,301 bytescontent/tracking.mdxasset21,868 bytescontent/turn-into-app.mdxasset14,022 bytescontent/updating-ui-in-production.mdxasset2,419 bytescontent/upgrading-core-ui.mdxasset21,943 bytescontent/using-your-agent.mdxasset8,303 bytescontent/vercel.mdxasset8,138 bytescontent/versioning-and-stability.mdxasset2,853 bytescontent/voice-input.mdxasset8,381 bytescontent/webmcp.mdxasset16,627 bytescontent/what-is-agent-native.mdxasset4,059 bytescontent/workspace-connections.mdxasset21,310 bytescontent/workspace-deployment.mdxasset12,511 bytescontent/workspace-management.mdxasset12,731 bytescontent/writing-agent-instructions.mdxasset11,943 bytesdesign/action-receipts-and-domain-events.mddoc2,597 bytesdesign/durable-agent-runs.mddoc63,763 bytesdesign/gpt-reasoning-effort-gateway-contract.mddoc4,588 bytesdesign/organization-team-tenancy.mddoc38,654 bytesdesign/provider-integration-authority.mddoc7,687 bytesmigrations/README.mddoc2,494 bytesmigrations/agentkit-chat.mddoc9,675 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing non-trivial Agent-Native functionality, read the
version-matched docs installed with `@agent-native/core`. When implementation
examples or template patterns matter, inspect the packaged source corpus too.
## How
1. Start from the generated app root.
2. Start cross-corpus discovery with `pnpm action framework-search --pattern "<feature>"`.
3. Use `--scope docs` or `--scope source` and `--mode glob`, `--mode sql-like`,
or `--mode regex` when the question needs a narrower pattern.View the complete source on GitHubauthentication
When your agent uses it
How to wire this app's login experience through the framework's createAuthPlugin — branded loginHtml, public paths, and signup/logout UI. Read before creating or editing server/plugins/auth.ts, or before adding gated routes / a public marketing page.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
2,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the landing route before wiring auth
- Branded login page
- Logging out
- Separating a public marketing page from the private app
- Public paths
- Example
Files
SKILL.mdcore11,623 bytesSKILL.mdcore11,623 bytes
Read SKILL.md
# Authentication
## Rule
If a user requests authentication or login or other similar terms, you should setup
authentication in their app using this skill.
`server/plugins/auth.ts` exports `createAuthPlugin(options)` from
`@agent-native/core/server` — that plugin is the app's auth. Do not hand-roll
session cookies, a parallel login route, or a custom auth backend; configure
the existing plugin instead.
## Choose the landing route before wiring auth
View the complete source on GitHubinternationalization
When your agent uses it
How to add localized UI copy when the user explicitly requests i18n. Do not load or apply this skill for ordinary English UI edits in this starter — it ships inline English strings with no catalogs.
What it loads into context
Frontmatter
63 tokensAlways available to the agent
SKILL.md
391 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Catalogs
- Verification
Files
SKILL.mdcore1,523 bytesSKILL.mdcore1,523 bytesSKILL.mdcore1,523 bytesSKILL.mdcore1,523 bytes
Read SKILL.md
# Internationalization
## Rule
**Opt-in only.** Do not add i18n catalogs, `LanguagePicker`, locale init scripts,
or `AppProviders i18n={{...}}` unless the user explicitly asks for
internationalization / localization / multiple languages.
This starter ships **English-only inline copy**. For normal UI text edits,
change the string in the component and stop — do not create `app/i18n/`.
When the user does request i18n, then: visible UI copy belongs in the app's i18n
catalog, not inline in components. Update the English source catalog first,
update existing locale catalogs, and run the i18n guard.View the complete source on GitHubmulti-app-workspace
When your agent uses it
How to build or extend a multi-app Agent-Native workspace — the monorepo shape, the Dispatch control plane, the shell app that fronts every other app, and the shared package. Use when the user asks for a workspace, a platform, a suite of apps, an app launcher or shell, a second app, Dispatch, a shared vault, or cross-app agents. This repo is a standalone app, so read this before restructuring it.
What it loads into context
Frontmatter
121 tokensAlways available to the agent
SKILL.md
3,508 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- This repo's starting point
- Scaffold
- Anatomy
- The shell / default app
- Dispatch
- Shared design system and brand
- Three layers, merged by file name
- packages/shared
- Don't
- Verify
- Deeper reading
Files
SKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytesSKILL.mdcore14,653 bytes
Read SKILL.md
# Multi-App Workspaces
## Rule
Scaffold a workspace with the CLI. Never hand-build the monorepo wiring, and
never hand-build Dispatch — it is a published template and runtime package, not
something to recreate from parts.
## This repo's starting point
`package.json` records `agent-native.scaffold.shape: "standalone"`, there is no
`agent-native.workspaceCore` field, and `pnpm-workspace.yaml` has no `packages:`
key. This is a single app, not a workspace root. `add-app` will not work here
until a workspace root exists.View the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubbuild-an-app
When your agent uses it
Fast end-to-end workflow for turning this starter into a working domain app from a short or vague request. Use when asked to build, make, create, or prototype an app, tool, tracker, dashboard, booking, inventory, or CRM system with Agent-Native. Read before the first edit.
What it loads into context
Frontmatter
87 tokensAlways available to the agent
SKILL.md
2,341 tokensWhen the skill is activated
Reference documents
11,894 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 1. Orient
- 2. Decide
- 3. Build: write everything, then verify once
- 4. Verify, once, in this order
- 5. Report
- Related
Files
SKILL.mdcore9,852 bytesSKILL.mdcore9,852 bytesSKILL.mdcore9,852 bytesreferences/data-and-access.mddoc7,061 bytesreferences/edit-points-chat.mddoc2,806 bytesreferences/edit-points-default.mddoc2,595 bytesreferences/rooms-example.mddoc33,575 bytesreferences/verification.mddoc8,399 bytes
Read SKILL.md
# Build an App
Done means one useful domain workflow with a designed route, shared UI/agent
actions, correct access, agent context, and a smoke run that proves its key
rule. Finish a narrow slice before expanding it.
This skill edits source files. If this session cannot write files (a deployed
app's runtime agent), say the change needs a code session and stop.
The four-area parity rule from `adding-a-feature` (UI, action, instructions,
application state) applies to everything you add. This skill supplies the
order and decisions; use the edit-point reference for the scaffold you have.
## 1. OrientView the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-engines
When your agent uses it
How to inspect and configure the AI engine (model provider) powering the agent. Use when the user asks to switch models, check which engine is active, test a new provider, or register a custom engine.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
2,883 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Overview
- Available Tools
- Checking the Current Engine
- Switching Engines
- Settings › Model
- Restricting Personal API Keys
- Testing a New Engine
- Choosing Which Models Show
- Built-in Engines
- Engine Capabilities
- Anthropic-Exclusive Features
- Using OpenRouter
- Registering a Custom Engine
- Env Vars Reference
Files
SKILL.mdcore11,634 bytes
Read SKILL.md
# Agent Engines
## Overview
The framework supports pluggable AI engines beneath the agent loop. The **Anthropic engine** is the default and best-in-class path (Claude models). Additional engines can be added via the Vercel AI SDK (OpenAI, Google Gemini, Groq, Mistral, Cohere, Ollama).
## Available Tools
| Tool | Purpose |
|---|---|
| `list-agent-engines` | List all registered engines, their capabilities, the current selection, and whether you can change the organization default (`canUpdateDefault`) |
| `set-agent-engine` | Set the organization's default engine and model (owners and admins only) |
| `test-agent-engine` | Send a trivial prompt to verify the engine works (connectivity + API key) |
| `check-provider-key` | Check a provider API key by listing the models it reaches, before saving it |View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubapp-branding
When your agent uses it
How to rename or rebrand the app so every surface — title, package metadata, PWA manifest, and icons — stays consistent. Use when a prompt establishes or implies a new app name, title, or visual identity, OR when you generate an app title, theme, or icon yourself while scaffolding an app from an initial prompt.
What it loads into context
Frontmatter
94 tokensAlways available to the agent
SKILL.md
1,749 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Trigger
- Rule
- Locate the layout first
- The Checklist
- What NOT to change
- Verification
Files
SKILL.mdcore6,872 bytesSKILL.mdcore6,872 bytesSKILL.mdcore6,872 bytes
Read SKILL.md
# App Branding
## Trigger
Read this skill whenever the app's name, title, or visual identity is being
established or changed — whether the user states it or **you derive it yourself**.
- **Explicit** — the prompt names the app: "rename this app to …", "this is a
Todo app called …", "call this app …".
- **Implicit / agent-generated** — the prompt describes what to build but leaves
the branding to you: "build me a habit tracker", "turn this into a recipe
manager", "make a CRM for my studio". The moment you decide on a title, theme
color, or icon from that prompt, treat it as a branding change and apply the
full checklist so your generated identity lands on every surface — not justView the complete source on GitHubapp-permissions
When your agent uses it
How to enable browser feature permissions (camera, microphone, geolocation, screen capture, clipboard, wake-lock) that the framework's default Permissions-Policy header blocks. Use when the app must access a device or browser capability that is silently failing.
What it loads into context
Frontmatter
75 tokensAlways available to the agent
SKILL.md
1,634 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Common directives
- Don't
- Related Skills
Files
SKILL.mdcore6,916 bytesSKILL.mdcore6,916 bytesSKILL.mdcore6,916 bytes
Read SKILL.md
# App Permissions — Enabling Browser Features
## Rule
When the app needs a browser feature that the default `Permissions-Policy`
disables, re-enable it for first-party use in
`server/plugins/permission-policy.ts` — change only the directive the feature
needs, never weaken the others, and never edit the `@agent-native/*` framework
package.
## Why
There are **two independent layers** between the app and a device capability, and
this skill only touches the first one:View the complete source on GitHubbuild-an-app
When your agent uses it
Fast end-to-end workflow for turning this starter into a working domain app from a short or vague request. Use when asked to build, make, create, or prototype an app, tool, tracker, dashboard, booking, inventory, or CRM system with Agent-Native. Read before the first edit.
What it loads into context
Frontmatter
87 tokensAlways available to the agent
SKILL.md
2,341 tokensWhen the skill is activated
Reference documents
11,894 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 1. Orient
- 2. Decide
- 3. Build: write everything, then verify once
- 4. Verify, once, in this order
- 5. Report
- Related
Files
SKILL.mdcore9,852 bytesSKILL.mdcore9,852 bytesSKILL.mdcore9,852 bytesreferences/data-and-access.mddoc7,061 bytesreferences/edit-points-chat.mddoc2,806 bytesreferences/edit-points-default.mddoc2,595 bytesreferences/rooms-example.mddoc33,575 bytesreferences/verification.mddoc8,399 bytes
Read SKILL.md
# Build an App
Done means one useful domain workflow with a designed route, shared UI/agent
actions, correct access, agent context, and a smoke run that proves its key
rule. Finish a narrow slice before expanding it.
This skill edits source files. If this session cannot write files (a deployed
app's runtime agent), say the change needs a code session and stop.
The four-area parity rule from `adding-a-feature` (UI, action, instructions,
application state) applies to everything you add. This skill supplies the
order and decisions; use the edit-point reference for the scaffold you have.
## 1. OrientView the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubinline-embeds
When your agent uses it
How to render interactive previews of app screens inline in agent chat via sandboxed iframes. Covers the `embed` fence syntax, the `postNavigate` pop-out helper, and when to prefer an embed over prose / a static image.
What it loads into context
Frontmatter
65 tokensAlways available to the agent
SKILL.md
816 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- The `embed` fence
- When to reach for an embed
- "Open in main window" button — `postNavigate`
- Adding an embed route to a template
Files
SKILL.mdcore3,484 bytes
Read SKILL.md
# Inline Embeds
The chat renderer lets you embed a sandboxed iframe pointing at any same-origin
route of this app, rendered inline with your reply. Use this when a live,
interactive preview communicates more than prose or a static screenshot — a
chart, an email thread, a slide, a form, a ticket.
## The `embed` fence
Emit a fenced code block with the language `embed`:
````
```embed
src: /email?id=msg_123View the complete source on GitHubnotifications
When your agent uses it
In-app notifications primitive with pluggable server-side channels. Use when the agent needs to surface progress, alerts, or completions to the user — both in-app (bell + toast) and out-of-band (webhook, Slack, custom).
What it loads into context
Frontmatter
62 tokensAlways available to the agent
SKILL.md
899 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Tools
- Sending
- Delivery
- Built-in Channels
- Registering a Custom Channel
- HTTP API
- UI
- Related
Files
SKILL.mdcore3,860 bytes
Read SKILL.md
# Notifications
`notify()` is the framework's "tell the user something" primitive. Every call persists a row to the inbox (drives the bell UI) and fans out to any registered server-side channels. Channels follow the same pluggable-provider pattern as `tracking` — register at startup, `notify()` fans out, errors are isolated.
Use for: *agent progress milestones, automation triggers firing, background job completions, critical errors*. Don't use for chat replies — those go through the conversation.
## Tools
| Tool | Purpose |
|---|---|
| `notify` | Send a notification (severity + title + optional body/metadata/channels) |
| `list-notifications` | Show recent notifications for the current user |
## SendingView the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubprogress
When your agent uses it
Report live progress from long-running agent tasks. Use when a task takes more than a few seconds, so the user can watch status in the runs tray instead of staring at a spinner.
What it loads into context
Frontmatter
50 tokensAlways available to the agent
SKILL.md
1,019 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Overview
- Tool
- Canonical Flow
- Best Practices
- Runs API
- UI Surface
- Event Bus Integration
- Related Skills
Files
SKILL.mdcore4,070 bytes
Read SKILL.md
# Progress
## Overview
`progress_runs` is the framework's "what is the agent doing right now" primitive. The agent starts a run at the top of a long task, updates it as work proceeds, and completes it with a terminal status. The UI renders active runs in a header-bar widget with a percent bar, current step, and spinner/check/X — live visibility into work that would otherwise be opaque.
Separate concern from `notifications`:
| | Notifications | Progress |
|---|---|---|
| Shape | One-shot event — "X happened" | Continuous state — "X is 45% done" |
| UI surface | Bell + toast | Runs tray with percent bar |
| Lifecycle | Dismissable (read/unread) | Running → terminal (succeeded/failed/cancelled) |
View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubreview-latest-feedback
When your agent uses it
Sweep recent feedback (the Slack feedback channels, GitHub issues, CI on main, Sentry, Analytics errors, chat run health, deploys), fix as many real bugs as possible, ship them in one PR, then close the loop on every Slack report with 👀/✅/🎫 and one short plain-English reply. Use for scheduled or manual feedback sweeps, usually followed by /ship.
What it loads into context
Frontmatter
119 tokensAlways available to the agent
SKILL.md
5,830 tokensWhen the skill is activated
Reference documents
4,243 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- The contract
- Boundaries
- 1. Collect (keep it under ~15 minutes)
- 2. Triage: lean toward fixing
- 3. Fix as many as you can
- 4. Ship
- 5. Close the loop
- 6. Recap
Files
SKILL.mdcore24,362 bytesSKILL.mdcore24,362 bytesSKILL.mdcore24,362 bytesSKILL.mdcore24,362 bytesSKILL.mdcore24,362 bytesSKILL.mdcore24,362 bytesreferences/slack-reaction-replies.mddoc19,339 bytes
Read SKILL.md
# Review Latest Feedback
The job: find what's broken, fix as much of it as you can, ship it, and tell
each reporter in one friendly sentence what happened. A sweep is judged by
bugs fixed and merged, and by every claimed report getting a clear answer. It is
not judged by ledgers, receipts, or checker output, so don't produce them.
## The contract
The detailed Slack reaction and reply rules are in
[slack-reaction-replies.md](references/slack-reaction-replies.md); read it before
marking, replying to, or closing any Slack report.
Slack reactions on a report's root message mean exactly this:View the complete source on GitHubreview-prs
When your agent uses it
Review recent BuilderIO/agent-native human pull requests, approve eligible internal PRs, and identify PRs ready for Steve's final maintainer decision. Outside Factory governance, merge only when Steve explicitly requests it for that PR. Factory governance is approval-only. Use for scheduled or manual PR sweeps; flag external work needing a major product call or new framework API, and unresolved major product or UX decisions, to Steve.
What it loads into context
Frontmatter
120 tokensAlways available to the agent
SKILL.md
9,277 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Selection and evidence
- Explicit blanket approval instruction
- External-contribution scope gate
- Liamdebeasi approval policy
- Internal-author approval policy
- Owner exceptions
- Standard approval gate
- Product and UX decisions
- Review actions
- Merge-ready disposition and handoff
- Worktrees and PR provenance
- End-of-run recap
- Related skills
Scripts executed separately
../../../../../../../../scripts/agent-friction-report.mjs../../../../../../../../scripts/agent-friction-report.mjs../../../../../../../../scripts/agent-friction-report.mjs../../../../../../../../scripts/agent-friction-report.mjs../../../../../../../../scripts/ci-change-scope.ts
Files
SKILL.mdcore43,975 bytesSKILL.mdcore43,975 bytesSKILL.mdcore43,975 bytesSKILL.mdcore43,975 bytesSKILL.mdcore43,975 bytesSKILL.mdcore43,975 bytesSKILL.mdcore43,975 bytesSKILL.mdcore43,975 bytesSKILL.mdcore43,975 bytesSKILL.mdcore43,975 bytesSKILL.mdcore43,975 bytes../../../../../../../../scripts/agent-friction-report.mjsruns230,778 bytes../../../../../../../../scripts/agent-friction-report.mjsruns230,778 bytes../../../../../../../../scripts/agent-friction-report.mjsruns230,778 bytes../../../../../../../../scripts/agent-friction-report.mjsruns230,778 bytes../../../../../../../../scripts/ci-change-scope.tsruns97,823 bytes
Read SKILL.md
# Review Pull Requests
Review the newest relevant human pull requests in `BuilderIO/agent-native`.
Under the default internal-author policy below, approve eligible fixes from
verified BuilderIO organization members. Never submit an approval review for an
external or unverified author; `community-prs` assigns that code-review
decision to a human Builder dev. When Steve explicitly gives the manual blanket
instruction for a non-Factory review sweep defined below, follow it over
ordinary approval gates; the non-waivable safeguards still apply.
Approval and merge readiness are separate. **Ready for your final approval**
means implementation is complete, required checks have passed, and any required
product or UX decision is complete. It identifies a candidate for Steve's
final decision; readiness alone does not authorize a merge. Factory governanceView the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHuba2a-protocol
When your agent uses it
How agent-native apps discover, authenticate with, and call each other over A2A. Use when connecting one app to another, debugging a remote agent that won't authenticate, exposing skills to peers, or calling agents from scripts.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
5,499 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- A2A is already mounted
- Registering a remote agent
- Authentication
- Hosted providers
- Advertising what this agent can do
- Calling another agent
- JSON-RPC Methods
- Task Lifecycle
- Message Parts
- Custom mount
- All Types
- Related Skills
Files
SKILL.mdcore23,694 bytesSKILL.mdcore23,694 bytes
Read SKILL.md
# A2A Protocol (Agent-to-Agent)
## Rule
Agents call other agents over A2A, a JSON-RPC protocol for discovery and
delegation. Use it only when the user explicitly asks to involve a peer, or
the requested outcome requires data, credentials, or a capability owned by
that peer and unavailable through the local app's actions. The local app's
actions are the default; a peer's availability or ability to enrich a result
is not enough. When a peer supplies input, keep authoring and editing artifacts
native to the local app here.
When a handoff qualifies under the rule above, preserve the peer boundary: do
not have app A generate and execute raw SQL against app B's database, or exposeView the complete source on GitHubactions
When your agent uses it
How to create and run agent actions. Actions are the single source of truth for app operations — the agent calls them as tools and frontend code calls them through client hooks. Use when creating a new action, adding an API integration, or wiring up frontend data fetching.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
4,670 tokensWhen the skill is activated
Reference documents
5,535 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Keep Actions Deterministic
- How to Create an Action
- Keep the Action Surface Small and Orthogonal
- Key Actions — One Index, Every Surface
- The `http` Option
- Return Values
- Frontend Hooks
- How to Run (Agent)
- Custom `/api/` Routes
- Do / Don't
- Troubleshooting
- References
- Related Skills
Files
SKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesSKILL.mdcore20,285 bytesreferences/action-fields.mddoc15,118 bytesreferences/examples.mddoc3,121 bytesreferences/provider-apis.mddoc6,265 bytes
Read SKILL.md
# Agent Actions
## Rule
Actions in `actions/` are the **single source of truth** for app operations. The agent calls them as tools, the frontend calls them through `useActionQuery` / `useActionMutation`, and the framework owns the HTTP transport behind those hooks — no duplicate `/api/` routes.
Before creating any custom route for app data, check `actions/` and the action table in `AGENTS.md`. An action already exists? Call it directly. Missing? Create or update a `defineAction`. **Stop trigger:** about to add a file under `server/routes/api/` (or middleware to guard one)? Check it against the exception list in *Custom `/api/` Routes* below first — even if you already started the route.
## Keep Actions Deterministic
An action may call a provider API, validate data, and persist records without being an AI feature — keep it deterministic, focused, and independently useful to the agent. Don't put LLM calls or a second model runtime in app actions.
The sanctioned model-backed exception is `@agent-native/core/decisions` for capability-aware classification and scoring, and it returns a result without performing or authorizing the selected action.
View the complete source on GitHubadding-a-feature
When your agent uses it
The four-area checklist every new feature must complete. Use when adding any feature, integration, or capability to ensure the agent and UI stay in parity.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
3,692 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Checklist
- Examples
- Adding a new route
- Anti-Patterns
- Verification
- One more area — sharing
- One more area — who inside the app may do it
- Related Skills
Files
SKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytesSKILL.mdcore16,465 bytes
Read SKILL.md
# Adding a Feature — The Four-Area Checklist
## Rule
Every new feature MUST update all four areas. Skipping any one breaks the agent-native contract — the agent and UI must always be equal partners.
Scope check first: this is the checklist for a feature that adds app data or a new capability. A restyle, a copy change, or a screen with no new data model and no new agent-visible operation only needs the areas it actually touches — do not manufacture actions, state keys, or a skill for it.
## Why
Agent-Native apps are defined by parity: everything the UI can do, the agent can do, and vice versa. A feature that only has UI is invisible to the agent. A feature that only has scripts is invisible to the user. A feature without app-state sync means the agent is blind to what the user is doing.
## The Checklist
View the complete source on GitHubadding-workspace-apps
When your agent uses it
How to add an app to an Agent-Native workspace: classifying an "agent" request, scaffolding `apps/<app-name>`, app discovery and descriptions, mounting and base paths, action-first data, and finishing a chat-template app. Use when asked to create, build, scaffold, or generate a new agent or app in a workspace.
What it loads into context
Frontmatter
92 tokensAlways available to the agent
SKILL.md
1,383 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Classify The Ask First
- Where The App Lives
- Vault Access
- Discovery, Links, And The UI Stack
- App Data Goes Through Actions
- PostgreSQL Database
- Scaffolding
- Finishing A Chat-Template App
- Related Skills
Files
SKILL.mdcore6,068 bytesSKILL.mdcore6,068 bytes
Read SKILL.md
# Adding Workspace Apps
## Classify The Ask First
When a user asks from Dispatch chat or through `@agent-native` in Slack or
Telegram to create, build, make, scaffold, or generate an "agent", classify
the ask first.
Simple Dispatch-native behavior such as a reminder, digest, monitor, routing
rule, saved instruction, or recurring workflow can stay in Dispatch as a
recurring job/resource/destination. Robust unique products or teammates that
need their own UI, data model, actions, integrations, or domain workflow should
become a separate workspace app under `apps/<app-name>`, mounted at
`/<app-name>`.
View the complete source on GitHubagent-native-docs
When your agent uses it
How to find version-matched Agent-Native framework docs and source bundled in node_modules. Use before implementing or answering questions about @agent-native/core APIs, generated apps, workspaces, templates, or advanced features.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
1,640 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Reuse Proven Patterns (rg + cp)
- Useful Slugs
- Don't
- Authoring visual docs blocks
Files
SKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytesSKILL.mdcore7,162 bytes
Read SKILL.md
# Agent-Native Docs Lookup
## Rule
Before implementing or explaining non-trivial Agent-Native behavior, read the
version-matched docs installed with `@agent-native/core`. When examples,
imports, or implementation details matter, inspect the packaged source corpus
too.
## Why
Generated apps and workspaces may be on a different framework version than the
public docs or model memory. The installed package is the source that matches
the app in front of you. A version-matched `@agent-native/core-corpus` packageView the complete source on GitHubagent-native-toolkit
When your agent uses it
Inventory and ownership rules for shared Agent-Native workspace UI. Use before building app chrome, settings, navigation, sharing, collaboration, setup, history, comments, chat rails, agent UX, or repeated workspace behavior.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
3,733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Core Rule
- Agent Surface Contract
- Visual Direction And Workspace Variety
- Discover Before Building
- Design-System Boundary
- Settings Direction
- Integration Setup Preflight
- Reusable Kits
- Implementation Checklist
- Related Skills
Files
SKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytesSKILL.mdcore17,085 bytes
Read SKILL.md
# Agent-Native Toolkit
Use this skill when deciding whether app chrome, settings, collaboration,
sharing, navigation, organization, setup, history, comments, or agent UX should
be built app-locally or moved into reusable framework/toolkit pieces.
## Core Rule
Apps own domain models, domain actions, and product-specific workflows. The
framework and `@agent-native/toolkit` own repeated workspace behavior users
expect to work the same everywhere.
Move behavior into shared toolkit primitives when it is:
View the complete source on GitHubautomations
When your agent uses it
Manual, event-triggered, webhook-triggered, and schedule-triggered automations with optional event conditions. Use when creating automations, wiring events or webhooks, running one on demand, or understanding triggers.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
4,449 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- The Persisted Trigger Types
- How It Works
- Markdown Format
- Agent Tools
- Organization Event Automations
- The Event Bus
- Condition Evaluator
- The `web-request` Tool and Keys
- Failure Handling
- UI
- Example
- Key Files
- Related Skills
Files
SKILL.mdcore20,286 bytes
Read SKILL.md
# Automations
## Rule
Automations are the user-facing umbrella for agent-executed tasks that run on
demand, in response to events or webhooks, or on a cron schedule. **Schedule**,
**Event**, **Manual**, and **Webhook** are the persisted trigger types. Each
automation is a markdown resource under `jobs/` with YAML frontmatter describing
when and how it fires, and a body containing natural-language instructions the
agent follows.
Recurring Jobs is the legacy name and API for scheduled automations.
`manage-jobs`, `jobs/`, and `/agent#jobs` remain stable compatibility surfaces.
Use `manage-automations` for new personal or organization automations, includingView the complete source on GitHubbuild-an-app
When your agent uses it
Fast end-to-end workflow for turning this starter into a working domain app from a short or vague request. Use when asked to build, make, create, or prototype an app, tool, tracker, dashboard, booking, inventory, or CRM system with Agent-Native. Read before the first edit.
What it loads into context
Frontmatter
87 tokensAlways available to the agent
SKILL.md
2,341 tokensWhen the skill is activated
Reference documents
11,894 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 1. Orient
- 2. Decide
- 3. Build: write everything, then verify once
- 4. Verify, once, in this order
- 5. Report
- Related
Files
SKILL.mdcore9,852 bytesSKILL.mdcore9,852 bytesSKILL.mdcore9,852 bytesreferences/data-and-access.mddoc7,061 bytesreferences/edit-points-chat.mddoc2,806 bytesreferences/edit-points-default.mddoc2,595 bytesreferences/rooms-example.mddoc33,575 bytesreferences/verification.mddoc8,399 bytes
Read SKILL.md
# Build an App
Done means one useful domain workflow with a designed route, shared UI/agent
actions, correct access, agent context, and a smoke run that proves its key
rule. Finish a narrow slice before expanding it.
This skill edits source files. If this session cannot write files (a deployed
app's runtime agent), say the change needs a code session and stop.
The four-area parity rule from `adding-a-feature` (UI, action, instructions,
application state) applies to everything you add. This skill supplies the
order and decisions; use the edit-point reference for the scaffold you have.
## 1. OrientView the complete source on GitHubclient-side-routing
When your agent uses it
How to add routes without remounting the app shell. Use when adding a new route, fixing agent sidebar reloads on navigation, or choosing between `root.tsx` layout and pathless `_app.tsx` layout patterns.
What it loads into context
Frontmatter
72 tokensAlways available to the agent
SKILL.md
759 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Hard Rule
- Two Correct Patterns
- Anti-Pattern
- Chat-First Routes
- Adding a New Route
- Related Skills
Files
SKILL.mdcore3,156 bytes
Read SKILL.md
# Client-Side Routing
## Rule
All templates are single-page apps using React Router. **Navigation must not remount the app shell.** The agent sidebar, document tree, and any other persistent chrome must survive route changes.
## Why
If the shell unmounts on every navigation, the agent chat reconnects/reloads, destroying in-progress work and hammering the backend.
## Hard Rule
**The app shell (AgentSidebar + any top-level navigation) must be mounted ONCE, above the `<Outlet />`.** Never wrap each page in its own `<AppLayout>` / `<Layout>`. React sees a different component at the outlet position on each nav and unmounts the entire subtree.
View the complete source on GitHubcomposable-mini-apps
When your agent uses it
Build focused workspace apps and cross-app workflows, using A2A only when a task needs a peer-owned capability. Use when designing mini-app boundaries.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
1,715 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Shape
- Discovery And Invocation
- Artifact Handoff
- Provider APIs
- Example
- Don't
- Related Skills
Files
SKILL.mdcore7,490 bytesSKILL.mdcore7,490 bytes
Read SKILL.md
# Composable Mini-Apps
## Rule
Prefer many one-job apps in a workspace over one oversized app. A headless app
can own a provider, dataset, workflow, or specialist action surface without a
full UI. This architecture does not make every task a cross-app task: use the
current app's actions by default. Call a peer only when the user asks or the
outcome requires data, credentials, or a capability only that peer can provide.
A peer's availability or ability to enrich the result is not enough. The app
that owns an artifact remains responsible for authoring and editing it.
## Shape
View the complete source on GitHubcontext-awareness
When your agent uses it
How the agent knows what the user is looking at. Use when exposing UI state to the agent, implementing view-screen or navigate actions, wiring navigation state, or debugging agent context issues.
What it loads into context
Frontmatter
61 tokensAlways available to the agent
SKILL.md
3,456 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- The Core Patterns
- Jitter Prevention
- Gold-Standard Example: Mail Template
- Do
- Don't
- Related Skills
Files
SKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytesSKILL.mdcore14,849 bytes
Read SKILL.md
# Context Awareness
## Rule
Apps can sync semantic navigation state when the route changes by using the
shared route-state hook. When screen context is available, Core includes a
`<current-screen>` snapshot in each agent turn, so the agent can get basic
route context without first calling another tool. Apps can add concise
visible-record details to that snapshot and expose selection in tab-scoped
state.
## Why
Without context awareness, the agent is blind. It asks "which email?" when the user is staring at one. It cannot act on the current selection, cannot provide relevant suggestions, and cannot modify what the user sees. Context awareness is what makes the agent feel like a collaborator rather than a disconnected chatbot.View the complete source on GitHubcustomizing-agent-native
When your agent uses it
How to configure, compose, or eject Agent-Native features into app-owned code. Use when overriding shared components or integrations, customizing a template, adding UI to chat or headless apps, or inspecting package source.
What it loads into context
Frontmatter
68 tokensAlways available to the agent
SKILL.md
2,022 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Use A Company Design System
- Customize A Shared Feature
- Eject A Feature
- Find The Installed Implementation
- Preserve The Agent-Native Contract
- App Shapes
- After Ejecting
- Don't
- Related Skills
Files
SKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytesSKILL.mdcore9,580 bytes
Read SKILL.md
# Customizing Agent-Native
## Rule
Start with the app's registered design system and local UI adapters, then use
the public feature's props, semantic components, controller, product slots,
and callbacks. If those seams are not enough, use the eject CLI to transfer
the smallest supported feature into the app and make that copy app-owned.
Never edit `node_modules`, deep-import a private source file at runtime, or
patch an `@agent-native/*` package.
Use this order:
1. Set brand tokens with the build-time theme configuration when tokens areView the complete source on GitHubdelegate-to-agent
When your agent uses it
Route user-facing AI work through the app's agent chat. Use when adding an AI-powered button, workflow, research task, generation flow, or follow-up.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
517 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Start a visible agent turn
- Keep the boundary clear
- Related skills
Files
SKILL.mdcore2,353 bytes
Read SKILL.md
# Delegate AI Work to Agent Chat
## Rule
Use the agent chat for work that needs AI reasoning, research, generation, or
multi-step decisions. Keep app actions deterministic; let the agent orchestrate
them so the user can see and steer the work.
## Start a visible agent turn
Call `sendToAgentChat()` from the UI and open the AgentSidebar for user-initiated
work:
```tsView the complete source on GitHubexternal-agents
When your agent uses it
Connect external agents and MCP hosts (Claude, Claude Desktop, Claude Code, ChatGPT custom MCP apps, Codex, Cursor, Claude Cowork, VS Code GitHub Copilot, Goose, Postman, MCPJam) to an agent-native app over MCP, and round-trip artifacts back into the UI with MCP Apps and deep links. Use when adding an action's `link` builder or `mcpApp`, wiring the `/_agent-native/open` route, exposing an "ingest" action to MCP/A2A, or scaffolding apps from an external agent.
What it loads into context
Frontmatter
148 tokensAlways available to the agent
SKILL.md
6,857 tokensWhen the skill is activated
Reference documents
2,355 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- App-facing MCP connectors
- Provider API connectors
- How
- Do
- Don't
- Related Skills
- Blueprint installer
Files
SKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesSKILL.mdcore28,695 bytesreferences/mcp-apps-embedding.mddoc10,363 bytes
Read SKILL.md
# External Agents (MCP bridge + deep links)
## Rule
- **The connected model is the author.** Write content yourself with create/update tools; never delegate via `ask_app` or wait on an unreachable in-app form — see `DEFAULT_AGENT_NATIVE_MCP_INSTRUCTIONS`.
- **Result shape and key tools are in `actions`** (Return Values, Key Actions).
- **Artifacts fit in one call.** Up to 500,000 characters; return a `storing-data` handle.
- **One tab, one id.** WebMCP sends `X-Agent-Native-Browser-Tab`; state resolves to the driving tab (`context-awareness`).
An agent-native app is reachable by any MCP-compatible host. Keep
setup simple: for workspace or cross-app access, add one remote MCP connector:
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page
controls whether that single connector reaches all apps or only selected apps,
and Dispatch filters `list_apps`, `ask_app`, and `open_app` to the granted set.View the complete source on GitHubfrontend-design
When your agent uses it
Product-specific interface design for an Agent-Native app. Use when creating or changing any user-facing UI, including screenshot feedback, copy or density cleanup, settings, controls, or a visual redesign.
What it loads into context
Frontmatter
56 tokensAlways available to the agent
SKILL.md
988 tokensWhen the skill is activated
Reference documents
1,095 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Choose a direction
- Keep the starting view focused
- Use the app's design system
- Coordinate the UI with the agent
- Interaction Responsiveness
- Make interactions clear and accessible
- Review the result
Files
SKILL.mdcore4,835 bytesreferences/visual-direction.mddoc5,473 bytes
Read SKILL.md
# Frontend Design
Design around the user's task and the app's existing visual language. A
distinctive interface comes from a clear product point of view, not extra
decoration. Keep implementation accessible and responsive.
## Choose a direction
Before a substantial new surface or redesign, decide who is using it, what
they came to do, and what they should see or do next. Pick a visual direction
that fits the product's brand and domain. For routine changes, inspect and
follow the existing components, type, color tokens, and spacing.
If a screen has two unrelated primary jobs, split the workflows or states.View the complete source on GitHubperformance
When your agent uses it
Keep app loads and interactions responsive as data grows. Use when a page feels slow, jumps, or lags, or when changing a data read, list, polling path, startup path, or performance-sensitive UI.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
841 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Load only what the current view needs
- Keep repeated work cheap
- Protect startup and the first paint
- Keep interactions responsive
Files
SKILL.mdcore4,073 bytesSKILL.mdcore4,073 bytes
Read SKILL.md
# Performance
Find the cost that users actually wait on, then reduce it without weakening
the result. A slow screen may be waiting on the server, moving too much data,
serial reads, repeated computation, or browser rendering; measure before adding
caches or extra complexity.
## Load only what the current view needs
- Shape list reads around the fields the list displays. Keep large document
bodies, histories, rendered output, and other rarely viewed data on a detail
or on-demand path. Return a small preview when the list needs one.
- Paginate or window unbounded histories and activity feeds. Start with the
recent or visible range and load more as the user asks for it.View the complete source on GitHubportability
When your agent uses it
How to keep template code PostgreSQL-specific and hosting-agnostic. Use when defining schemas, writing raw SQL, or creating server routes.
What it loads into context
Frontmatter
49 tokensAlways available to the agent
SKILL.md
956 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Database
- Hosting Agnostic
- Related Skills
Files
SKILL.mdcore4,257 bytesSKILL.mdcore4,257 bytes
Read SKILL.md
# Portability
## Rule
Templates use one PostgreSQL schema and query contract. Local PGlite and hosted
PostgreSQL share the same SQL semantics.
## Database
Use Drizzle's PostgreSQL exports directly for schemas and its query builder for reads/writes:
```ts
import { sql } from "drizzle-orm";
import {View the complete source on GitHubreal-time-sync
When your agent uses it
Decide whether a page needs external changes before refresh, then use the opt-in shared SSE and polling transport safely.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
1,367 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Decision rule
- Cost and behavior
- Use the hook
- Query freshness
- Source counters
- Avoid
Files
SKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytesSKILL.mdcore6,165 bytes
Read SKILL.md
# Real-Time Sync
## Decision rule
Opt in only when both are true:
1. The data can change without the current user acting.
2. The user needs to see that change before refreshing or navigating.
A local agent edit does not meet this rule. The chat run stream already
invalidates action queries and source counters after each side-effecting tool
and again at run end, without opening a background poll.
### Opt inView the complete source on GitHubrecurring-jobs
When your agent uses it
The legacy manage-jobs compatibility surface, job files under jobs/, and cron scheduler internals. Use when maintaining manage-jobs integrations or debugging the scheduler.
What it loads into context
Frontmatter
52 tokensAlways available to the agent
SKILL.md
985 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Connected MCPs in background jobs
- Job Tool (built in)
- Job files an app owns
- UI Surface
- Key Files
- Related Skills
Files
SKILL.mdcore4,497 bytes
Read SKILL.md
# Recurring Jobs (Compatibility)
## Rule
Recurring Jobs is the legacy name and API for scheduled automations. Jobs live
as resource files under `jobs/` with YAML frontmatter for scheduling metadata.
Read `automations` for the product model, trigger types, frontmatter fields,
organization scope, and when to prefer `manage-automations` over `manage-jobs`.
This skill covers only what is specific to the legacy scheduled path: the
`manage-jobs` tool and the cron scheduler behind it.
## How It Works
View the complete source on GitHubreliable-mutations
When your agent uses it
How the agent must perform writes so they actually persist under the hosted foreground run budget and long-running background handoffs. Use whenever you create, update, delete, or batch-write app data — especially "do this for many items" loops, or any task where the user expects N things to end up saved.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
808 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- How
- Don't
- Related
Files
SKILL.mdcore3,447 bytes
Read SKILL.md
# Reliable Mutations
## Rule
Make a change in **one atomic call** when the action supports it, then **verify
the persisted end state and report concrete proof** (counts/ids). Never drive a
multi-step change by looping many small writes, and never report success from a
tool ✓ alone.
## Why
Hosted foreground agent runs have a short soft budget so they can hand off
cleanly under synchronous serverless walls. Durable background runs get a much
longer budget, but they still should not rely on loops of many small writes:View the complete source on GitHubsecrets
When your agent uses it
Connect external services and keep app credentials scoped. Use before adding an API key, OAuth connection, provider setup, or server-side request that needs a user's or workspace's credential.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
768 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Choose the credential source
- Register an app-local API key
- Resolve credentials on the server
- Related skills
Files
SKILL.mdcore3,613 bytesSKILL.mdcore3,613 bytes
Read SKILL.md
# Secrets
## Rule
Credentials belong to the user, organization, workspace, or deployment that
owns them. Reuse the strongest shared connection or credential flow available;
never copy a value into app code or client-visible state.
## Choose the credential source
1. Check the workspace connection catalog first. If the provider is already
connected, request the app grant and resolve its vault-backed credential
reference through the workspace-connection API.
2. For OAuth, use the shared authorization and token flow. Do not store accessView the complete source on GitHubsecurity
When your agent uses it
Secure app data, actions, routes, and external input. Use when changing authentication, authorization, database access, uploads, untrusted content, or server-side requests to user-provided URLs.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
800 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Validate input and queries
- Authorize rows and actions
- Handle untrusted content and URLs
- Protect credentials and consequential actions
- Related skills
Files
SKILL.mdcore3,734 bytes
Read SKILL.md
# Security
## Rule
Treat each action, route, query, and external request as a trust boundary.
Validate inputs, check the caller's authority for every row, and fail closed
when access cannot be established.
## Validate input and queries
- Define action inputs with a Zod `schema`; validate data from forms, URLs,
providers, and files before using it.
- Build database queries with Drizzle. For raw SQL, bind values as parameters;
never concatenate user input into SQL text.View the complete source on GitHubself-modifying-code
When your agent uses it
Tiers, checkpoints, and off-limits paths for an agent editing the app's own source. Use when designing UI for agent editability, deciding whether a file is safe for the agent to change, or when tempted to patch config, `.env`, or an `@agent-native/*` package. Do not load it for ordinary source edits.
What it loads into context
Frontmatter
98 tokensAlways available to the agent
SKILL.md
1,324 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Why
- Modification Taxonomy
- Git Checkpoint Pattern
- Designing for Agent Editability
- Don't
- Related Skills
Files
SKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytesSKILL.mdcore5,885 bytes
Read SKILL.md
# Self-Modifying Code
## Rule
The agent can edit the app's own source code — components, routes, styles, scripts. This is a feature, not a bug. Design your app expecting this.
## Why
An agent-native app isn't just an app the agent can _use_ — it's an app the agent can _change_. The agent can fix bugs, add features, adjust styles, and restructure code. This makes the agent a true collaborator, not just an operator.
## Modification Taxonomy
Not all modifications are equal. Use this to decide what level of care is needed:
View the complete source on GitHubshadcn-ui
When your agent uses it
Using the app's local UI adapter and Toolkit or shadcn primitives. Use when choosing or composing standard controls, implementing dialogs, menus, or forms, or changing component variants and theme tokens.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
733 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Keep imports behind the local adapter
- Compose existing controls
- Keep controls operable
- Related skill
Files
SKILL.mdcore3,486 bytes
Read SKILL.md
# shadcn/ui and Local UI Adapters
Use the components already provided by the app. Read the app's design-system
configuration, Toolkit provider, and local UI components before adding or
changing a control. Their exports, variants, and props are the contract; do
not guess from upstream examples.
## Keep imports behind the local adapter
Product pages and domain components should import standard controls through the
app's configured local UI alias. In the Chat starter, that is
`@/components/ui/*`; its `app/components/ui/button.tsx` is a one-line
re-export of `@agent-native/toolkit/ui/button`. Follow an adjacent local shim
when the app needs another Toolkit primitive that is already exported there.View the complete source on GitHubsharing
When your agent uses it
Framework-level sharing and privacy for user-authored resources (dashboards, documents, forms, decks, etc.). Use when making a resource table ownable, wiring list/read/update access checks, or dropping the standard share dialog into a template.
What it loads into context
Frontmatter
73 tokensAlways available to the agent
SKILL.md
3,345 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- Concepts
- Make a resource ownable
- Filter list/read queries
- Guard write actions
- Create actions must set owner
- Drop in the share UI
- Standard share surface
- Actions available everywhere
- Migration pattern for existing tables
- Templates that opt out
- Debugging
Files
SKILL.mdcore14,701 bytes
Read SKILL.md
# Sharing — Private by Default, Explicit Share
## Rule
Any resource a user **creates** (dashboards, documents, forms, decks, compositions, booking links, issues, analyses) is **private to the creator** by default and visible to others only when they have been **explicitly shared** with or when the creator changes visibility to `org` or `public`.
This is the framework-level primitive. Every ownable resource gets it for free — same API, same UI, same skill.
Workspace apps are the intentional organization-scoped exception: their default
visibility is `org`, configurable by an organization owner/admin. The creator is
stored as the app owner, and the creator or an organization owner/admin can
change the app's visibility or manage its share grants.
## ConceptsView the complete source on GitHubstoring-data
When your agent uses it
How to store application data in agent-native apps. All data lives in SQL. Use when adding data models, deciding where to store data, or reading/writing application data.
What it loads into context
Frontmatter
58 tokensAlways available to the agent
SKILL.md
4,181 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Rule
- How It Works
- Do
- Don't
- Security
- Application State and Context Awareness
- Related Skills
Files
SKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytesSKILL.mdcore18,800 bytes
Read SKILL.md
# Storing Data — SQL is the Source of Truth
## Rule
All application data lives in **SQL** (local PGlite, hosted Postgres in production). The agent and UI share the same database. SQL stores structured records, metadata, references, and searchable text — not large raw file payloads. Do not store durable app data in the filesystem unless the app is explicitly running a Local File Mode artifact flow described below.
Large binary or file-like payloads (images, video/audio, PDFs, ZIPs, screenshots, session replay chunks, thumbnails, generated assets, `data:` URLs, and base64 file bodies) must go through configured file/blob storage such as `uploadFile()` or `putPrivateBlob()`. Persist only the returned URL, asset id, or opaque blob handle in SQL. If storage is unavailable in hosted or persistent-database mode, fail closed with setup guidance instead of falling back to base64 in `application_state`, `settings`, `resources`, or app tables.
**Uploaded files an action opens later** (chat attachments, import controls, chunked commits) are one core `AttachmentRef`: mint with `mintAttachmentRef`, open with `resolveAttachment` and `unwrapAttachment` from `@agent-native/core/private-blob`. Never write a template descriptor, never branch on error text, and never truncate a ref (`ATTACHMENT_REF_MAX_CHARS` is the floor for any prompt or field that carries one). Each failure is typed: `notFound`, `forbiddenScope`, `expired` and `malformed` are definitive and stop the turn, except a value that is not a ref at all or is missing (`malformed` with `unrecognized_scheme` or `empty`), which is an ordinary action error the model corrects by passing the listed ref; only `storageUnavailable` is retryable, and the same ref works once storage is back, so never ask the user to attach the file again for it.
**Local File Mode exception:** some artifact apps (Content, Plans, Slides, Dashboards, Designs, etc.) can intentionally use repo files as the source of truth for the artifact itself. This must be explicit via `agent-native.json`, `AGENT_NATIVE_MODE=local-files`, or an app-owned local-file action helper. In that mode, the UI and agent still go through app actions, but those actions read/write scoped files through `@agent-native/core/local-artifacts` instead of SQL rows. App state, auth, settings, credentials, collaboration metadata, and hosted database mode remain SQL. File-to-database or file-to-provider synchronization is an explicit sync step, not an implicit side effect of editing.
When you add a data model, a list, or a read path, also follow the `performance` skill: project only the columns a list renders, index the columns hot queries filter/sort on, and avoid query waterfalls — so apps stay fast as data grows.
View the complete source on GitHubturn-into-app
When your agent uses it
Turns a thread, skill, spreadsheet, or Claude/ChatGPT project into a polished, visual Agent-Native app: populated domain screens with the in-app agent working behind contextual controls. Use when a user invokes `/turn-into-app`, or asks to turn a workflow, skill, spreadsheet, or project into an app, UI, workbench, or dashboard, including from Claude or ChatGPT on the web.
What it loads into context
Frontmatter
110 tokensAlways available to the agent
SKILL.md
3,995 tokensWhen the skill is activated
Reference documents
33,236 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Host, source, and when to ask
- Workflow
- Final report
- Related skills
Files
SKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesSKILL.mdcore16,932 bytesagents/openai.yamlasset383 bytesreferences/attachments.mddoc1,503 bytesreferences/fresh-project.mddoc4,733 bytesreferences/local-run-and-deploy.mddoc7,649 bytesreferences/review-loop.mddoc19,749 bytesreferences/source-brief.mddoc6,588 bytesreferences/spreadsheet-source.mddoc14,401 bytesreferences/ui-archetypes.mddoc19,180 bytesreferences/ui-direction.mddoc19,431 bytesreferences/ui-palettes.mddoc23,241 bytesreferences/where-it-runs.mddoc4,905 bytes
Read SKILL.md
# Turn Into App
Give a proven workflow an Agent-Native app face and agent brain. Start with
populated domain screens. Object controls send work to the agent; the object
shows progress and receives the result. Avoid form-under-stepper layouts.
## Host, source, and when to ask
### Classify the host
- Local scaffolding needs Node.js 22.22.0 or newer (with npm/npx) and pnpm.
- **Local host**: a terminal, a filesystem, and a target checkout (Claude Code,
Codex, Cursor, or an Agent-Native agent that can edit files). Build the app in
that checkout with steps 1-6, unless the route is Builder hosting (below).View the complete source on GitHubturn-into-skill
When your agent uses it
Capture a proven thread or workflow as a reusable Agent-Native skill and save it as a personal or organization resource. Use when a user invokes `/turn-into-skill` or asks to save a workflow for reuse.
What it loads into context
Frontmatter
78 tokensAlways available to the agent
SKILL.md
1,147 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Default behavior
- 1. Extract the repeatable workflow
- 2. Author the skill
- 3. Save at the requested scope
- 4. Verify and hand off
- Related skills
Files
SKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesSKILL.mdcore5,125 bytesagents/openai.yamlasset316 bytes
Read SKILL.md
# Turn Into Skill
This is a packaging-and-save workflow, not a request for a proposal. A skill
is the lightweight path for a repeatable job that mostly needs guidance and
existing app actions. Use `turn-into-app` when the job needs a dedicated UI,
new persistent domain state, or a distinct app route.
## Default behavior
- With no argument, use the current thread as the source.
- With a named skill, local workflow, transcript path, or attachment, read that
source and package it immediately.
- Do not ask the user to restate context already present in the source.
- Default to a personal skill unless the user explicitly asks for a team,View the complete source on GitHubworkspace-conventions
When your agent uses it
Shared conventions for a multi-app Agent-Native workspace: finding version-matched framework docs and source, shared vs app-owned code, file and blob storage, env and secrets, agent scratch files, and Dispatch Resources. Use when working across workspace apps, storing files or credentials, or looking up framework APIs.
What it loads into context
Frontmatter
82 tokensAlways available to the agent
SKILL.md
1,377 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Framework Docs Lookup
- Shared Conventions
- Related Skills
Files
SKILL.mdcore6,092 bytes
Read SKILL.md
# Workspace Conventions
The long-form detail behind the workspace `AGENTS.md` core rules. Read this
before adding shared code, storing files or secrets, creating workspace
resources, or looking up framework APIs.
## Framework Docs Lookup
Read `agent-native-docs` for the lookup workflow (`framework-search` first, then
the focused `docs-search` / `source-search` readers, and the `rg` fallback).
The workspace-specific slugs it does not list are `workspace` and
`multi-app-workspace`. Feature flags and other specialized workflows are
optional; read their skill only when the app uses them. Use package docs for framework APIs,View the complete source on GitHubcreative-context
When your agent uses it
Search and reuse approved brand examples before generating decks, designs, content, assets, or dashboards.
What it loads into context
Frontmatter
26 tokensAlways available to the agent
SKILL.md
923 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Before generation
- Reuse ladder
- Record the result
- Isolated deployment protocol
- Feedback and promotion
Files
SKILL.mdcore4,562 bytesSKILL.mdcore4,562 bytesSKILL.mdcore4,562 bytes
Read SKILL.md
# Creative Context
Use this skill whenever imported company context can influence a creative
output.
## Before generation
1. Read `contextMode`, `pinnedPackId`, and `selectedContextId` from the
`creative-context` application-state record. If the mode is `off`, do not
retrieve or apply brand DNA.
2. If a pack is pinned, load that exact immutable pack. Otherwise honor the
selected named context; when none is explicit, use Default plus at most one
app-bound or semantically matching specialty context.
3. Inspect the strongest two to five results with `get-context-item`; treat allView the complete source on GitHubWhen your agent uses it
What it loads into context
Frontmatter
0 tokensAlways available to the agent
SKILL.md
694 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Reading Annotations
- Actions
- Workflow
- Tips
- Adding Pinpoint to a New App
Files
SKILL.mdcore2,601 bytes
Read SKILL.md
# Pinpoint — Visual Feedback Tool
You are an agent with access to Pinpoint, a visual feedback and annotation tool. Users annotate UI elements on web pages and send structured feedback to you.
## Reading Annotations
Pins are stored as JSON files in `data/pins/{uuid}.json`:
```json
{
"id": "uuid",
"pageUrl": "/dashboard",
"comment": "This button color is wrong",
"element": {
"tagName": "button",View the complete source on GitHubavailability
When your agent uses it
How schedules, weekly rules, date overrides, travel schedules, and out-of-office entries combine to determine when someone is bookable.
What it loads into context
Frontmatter
32 tokensAlways available to the agent
SKILL.md
395 tokensWhen the skill is activated
Reference documents
0 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Schedule
- Weekly availability
- Date overrides
- Travel schedules
- Out of office
- Common tasks
Files
SKILL.mdcore1,557 bytes
Read SKILL.md
# Availability
## Schedule
A named set of rules (e.g. "Working Hours"). Each user has ≥1 schedule and
marks one as `isDefault`. Event types either use the user's default or
reference a specific schedule.
Timezone is set on the schedule, not the user — lets you have a "Europe
hours" schedule and a "US hours" schedule for the same person.
## Weekly availability
Rows in `schedule_availability`: (day 0-6, startTime "HH:MM", endTimeView the complete source on GitHubSkills included
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- external-agents
- frontend-design
- performance
- portability
- real-time-sync
- reliable-mutations
- secrets
- security
- self-modifying-code
- shadcn-ui
- sharing
- storing-data
- turn-into-app
- turn-into-skill
- workspace-conventions
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- external-agents
- frontend-design
- performance
- portability
- real-time-sync
- reliable-mutations
- secrets
- security
- self-modifying-code
- shadcn-ui
- sharing
- storing-data
- turn-into-app
- turn-into-skill
- workspace-conventions
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- external-agents
- frontend-design
- performance
- portability
- real-time-sync
- reliable-mutations
- secrets
- security
- self-modifying-code
- shadcn-ui
- sharing
- storing-data
- turn-into-app
- turn-into-skill
- workspace-conventions
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- external-agents
- frontend-design
- performance
- portability
- real-time-sync
- reliable-mutations
- secrets
- security
- self-modifying-code
- shadcn-ui
- sharing
- storing-data
- turn-into-app
- turn-into-skill
- workspace-conventions
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- external-agents
- frontend-design
- performance
- portability
- real-time-sync
- reliable-mutations
- secrets
- security
- self-modifying-code
- shadcn-ui
- sharing
- storing-data
- turn-into-app
- turn-into-skill
- workspace-conventions
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- external-agents
- frontend-design
- performance
- portability
- real-time-sync
- reliable-mutations
- secrets
- security
- self-modifying-code
- shadcn-ui
- sharing
- storing-data
- turn-into-app
- turn-into-skill
- workspace-conventions
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- external-agents
- frontend-design
- performance
- portability
- real-time-sync
- reliable-mutations
- secrets
- security
- self-modifying-code
- shadcn-ui
- sharing
- storing-data
- turn-into-app
- turn-into-skill
- workspace-conventions
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- external-agents
- frontend-design
- performance
- portability
- real-time-sync
- reliable-mutations
- secrets
- security
- self-modifying-code
- shadcn-ui
- sharing
- storing-data
- turn-into-app
- turn-into-skill
- workspace-conventions
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- external-agents
- frontend-design
- performance
- portability
- real-time-sync
- reliable-mutations
- secrets
- security
- self-modifying-code
- shadcn-ui
- sharing
- storing-data
- turn-into-app
- turn-into-skill
- workspace-conventions
- agent-native-docs
- authentication
- internationalization
- multi-app-workspace
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- build-an-app
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- frontend-design
- performance
- real-time-sync
- reliable-mutations
- secrets
- security
- shadcn-ui
- sharing
- storing-data
- actions
- adding-a-feature
- agent-engines
- agent-native-docs
- agent-native-toolkit
- app-branding
- app-permissions
- build-an-app
- client-side-routing
- context-awareness
- customizing-agent-native
- delegate-to-agent
- frontend-design
- inline-embeds
- notifications
- performance
- progress
- real-time-sync
- reliable-mutations
- secrets
- security
- shadcn-ui
- sharing
- storing-data
- review-latest-feedback
- review-prs
- actions
- adding-a-feature
- agent-native-docs
- agent-native-toolkit
- customizing-agent-native
- delegate-to-agent
- performance
- reliable-mutations
- secrets
- security
- self-modifying-code
- sharing
- storing-data
- a2a-protocol
- actions
- adding-a-feature
- adding-workspace-apps
- agent-native-docs
- agent-native-toolkit
- automations
- build-an-app
- client-side-routing
- composable-mini-apps
- context-awareness
- customizing-agent-native
- delegate-to-agent
- external-agents
- frontend-design
- performance
- portability
- real-time-sync
- recurring-jobs
- reliable-mutations
- secrets
- security
- self-modifying-code
- shadcn-ui
- sharing
- storing-data
- turn-into-app
- turn-into-skill
- workspace-conventions
- creative-context
- availability
Showing 300 of 848 skills. See GitHub for the complete repository.
Repository facts
- License
- MIT
- Created
- Mar 12, 2026
- Last pushed
- Nov 7, 2026
- Forks
- 699
- Open issues
- 76
- Project activity
- Active
Rank over the last 12 weeks
Now #192 · best #183 · down 3 in 30 days
- #188 on Oct 29
- #189 on Oct 30
- #190 on Oct 31
- #189 from Nov 1 to Nov 2
- #191 from Nov 3 to Nov 4
- #192 since Nov 5
73 earlier periods
- #213 on Aug 2
- #214 on Aug 8
- #216 on Aug 9
- #222 on Aug 10
- #221 on Aug 11
- #220 on Aug 12
- #224 from Aug 13 to Aug 14
- #223 on Aug 15
- #222 from Aug 16 to Aug 17
- #225 from Aug 18 to Aug 19
- #222 on Aug 20
- #224 on Aug 21
- #222 on Aug 22
- #229 on Aug 23
- #228 on Aug 24
- #229 on Aug 25
- #230 on Aug 26
- #231 on Aug 27
- #229 on Aug 28
- #228 on Aug 29
- #227 on Aug 30
- #229 on Aug 31
- #233 on Sep 1
- #228 on Sep 2
- #235 on Sep 3
- #231 on Sep 4
- #234 on Sep 5
- #231 on Sep 6
- #233 from Sep 7 to Sep 8
- #234 on Sep 9
- #232 on Sep 10
- #233 from Sep 11 to Sep 12
- #236 on Sep 13
- #237 on Sep 14
- #238 on Sep 15
- #236 on Sep 16
- #235 on Sep 17
- #238 on Sep 18
- #234 on Sep 19
- #230 on Sep 20
- #207 on Sep 21
- #199 on Sep 22
- #194 on Sep 23
- #190 from Sep 24 to Sep 26
- #191 on Sep 27
- #192 on Sep 28
- #191 on Sep 29
- #188 on Sep 30
- #190 on Oct 1
- #191 on Oct 2
- #188 on Oct 3
- #185 on Oct 4
- #188 on Oct 5
- #189 on Oct 6
- #191 on Oct 7
- #189 on Oct 8
- #191 on Oct 9
- #189 on Oct 10
- #190 on Oct 11
- #191 on Oct 12
- #193 on Oct 13
- #192 on Oct 14
- #185 on Oct 15
- #183 on Oct 16
- #185 on Oct 17
- #186 from Oct 18 to Oct 19
- #185 on Oct 20
- #187 on Oct 21
- #186 on Oct 22
- #185 on Oct 23
- #186 on Oct 24
- #187 on Oct 25
- #187 on Oct 28
Before you install
Network or broad shellAutomated signals, not a guarantee
Checked Nov 2, 2026
Changed on Oct 27, 2026
No longer found: Dangerous shell operation, Possible obfuscation, Access to secrets
Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.
Show 3,227 notes
.agents/plugins/agent-native-turn-into-app/.claude-plugin/plugin.json:1Analysis incomplete: Referenced executable target unavailable; not analyzed.claude-plugin/marketplace.json:1Analysis incomplete: Referenced executable target unavailable; not analyzed.codex/hooks.json:1Analysis incomplete: Referenced executable target unavailable; not analyzed.codex/hooks.json:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/actions/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/actions/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/actions/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/actions/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/adding-a-feature/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/adding-a-feature/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/adding-a-feature/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/agent-native-docs/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/agent-native-docs/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/agent-native-docs/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/agent-native-docs/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/agent-native-docs/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/agent-native-docs/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/agent-native-docs/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/agent-native-docs/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedcommunity-templates/account-expert/.agents/skills/agent-native-docs/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzed
- No broad tool permissions
- 67 dangerous shell commandsin 34 files: docs/content/locales/ar-SA/security.mdx:111 docs/content/locales/de-DE/security.mdx:111 docs/content/locales/es-ES/security.mdx:111 and 31 more
- 6,735 network referencesin 1,332 files: mail-interaction-parity/references/interaction-matrix.md:25 scripts/agent-friction-report.mjs:1259 .agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:18 and 1,329 more
- 38 obfuscation signalsin 32 files: docs/content/locales/ko-KR/external-agents.mdx:230 docs/content/locales/ko-KR/mcp-apps.mdx:37 docs/content/locales/pt-BR/mcp-clients.mdx:23 and 29 more
- 4 instruction override signalsin 4 files: docs/content/locales/es-ES/pr-visual-recap.mdx:429 docs/content/locales/ja-JP/pr-visual-recap.mdx:439 docs/content/locales/pt-BR/pr-visual-recap.mdx:535 and 1 more
- 10 executable scriptsin 2 files: scripts/agent-friction-report.mjs:1 scripts/ci-change-scope.ts:1
- 1,724 references to secretsin 631 files: docs/content/environment-variables.mdx:4 agent-engines/SKILL.md:47 agent-engines/SKILL.md:47 and 628 more
- No requests to vote in Hot
Show all 8,578 signals by file
.agents/plugins/agent-native-design/.claude-plugin/plugin.json
- Network accessline 5
"author": { "name": "Agent-Native", "url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" - Network accessline 6
"name": "Agent-Native", "url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" }, - Network accessline 7
"url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" - Network accessline 8
}, "homepage": "https://design.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-homepage", "repository": "https://github.com/BuilderIO/agent-native", - Network accessline 9
"homepage": "https://design.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-homepage", - Network accessline 10
"repository": "https://github.com/BuilderIO/agent-native",
.agents/plugins/agent-native-design/.mcp.json
.agents/plugins/agent-native-design/skills/design-exploration/SKILL.md
- Network accessline 95
- Hosted default: connect `https://design.agent-native.com/mcp`. - Network accessline 108
the user the reconnect step: in Claude Code run `/mcp` and choose Authenticate/Reconnect for the Design connector; from any terminal run `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` — this - Network accessline 109
Authenticate/Reconnect for the Design connector; from any terminal run `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` — this re-authenticates WITHOUT reinstalling. Never reinstall from scratch just to fix - Network accessline 110
`npx -y @agent-native/core@latest reconnect https://design.agent-native.com` — this - Network accessline 111
re-authenticates WITHOUT reinstalling. Never reinstall from scratch just to fix auth. Continue once the connector is available. - Do not hand-roll MCP HTTP requests with curl from the agent session. Use the - Network accessline 112
auth. Continue once the connector is available. - Do not hand-roll MCP HTTP requests with curl from the agent session. Use the host-exposed Design tools after restart/reload, or use the returned - Network accessline 113
- Do not hand-roll MCP HTTP requests with curl from the agent session. Use the
.agents/plugins/agent-native-design/skills/visual-edit/README.md
- Network accessline 55
See the [skill instructions](https://github.com/BuilderIO/agent-native/blob/main/skills/visual-edit/SKILL.md)
.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md
- Network accessline 18
The editor is hosted at `https://design.agent-native.com`. Never start local - Network accessline 21
See the [Visual Edit guide](https://github.com/BuilderIO/agent-native/blob/main/skills/visual-edit/README.md) - Network accessline 57
- Inline-browser hosts should open `https://design.agent-native.com/visual-edit` - Network accessline 73
JavaScript evaluator. Prefer the host's inline browser; use external Chrome only when requested or unavailable. Open `https://design.agent-native.com/visual-edit`, keep it visible while tools - Network accessline 74
only when requested or unavailable. Open `https://design.agent-native.com/visual-edit`, keep it visible while tools register, and read the title before work. Never enter, copy, or request - Network accessline 75
`https://design.agent-native.com/visual-edit`, keep it visible while tools - Network accessline 78
other pages that block tools behind sign-in still require a signed-in tab. If the browser exposes CDP permissions, grant `local-network-access` to `https://design.agent-native.com` before calling page tools; otherwise use the - Network accessline 79
If the browser exposes CDP permissions, grant `local-network-access` to `https://design.agent-native.com` before calling page tools; otherwise use the page's Connect button and let the browser's permission prompt complete. - Network accessline 80
`https://design.agent-native.com` before calling page tools; otherwise use the - Network accessline 99
} const result = await an.call("open-visual-edit", { devServerUrl: "http://localhost:5173", - Network accessline 100
const result = await an.call("open-visual-edit", { devServerUrl: "http://localhost:5173", paths: ["/"], - Network accessline 101
devServerUrl: "http://localhost:5173", - Network accessline 175
- Browser hosts can use page-local `get-visual-edit-prompt`. - The `open-visual-edit` action is owned by Design. From another app, use the hosted MCP server at `https://design.agent-native.com/mcp` or the page's - Network accessline 176
- The `open-visual-edit` action is owned by Design. From another app, use the hosted MCP server at `https://design.agent-native.com/mcp` or the page's WebMCP helper, not `pnpm action` in the target app. The page path works - Network accessline 177
hosted MCP server at `https://design.agent-native.com/mcp` or the page's - Network accessline 198
path/query for another state; preserve the order of named or numbered flows. Shorthand like `localhost:1234/onboarding/1` means `http://localhost:1234/onboarding/1`. - Network accessline 200
`http://localhost:1234/onboarding/1`. - Network accessline 279
[Credential-related source omitted] - Network accessline 280
[Credential-related source omitted] - Access to secretsline 280
[Credential-related source omitted] - Network accessline 291
npx @agent-native/core@latest design connect --url http://localhost:5173 --root . --json - Network accessline 303
pnpm dev:cli design connect --url http://localhost:5173 --root templates/<app> --json - Network accessline 315
[Credential-related source omitted] - Network accessline 318
(Equivalently, pass `--bridge-token <token>`.) This starts a detached bridge on `http://127.0.0.1:7331`, adopts the server-minted token — so bridge and row - Network accessline 319
`http://127.0.0.1:7331`, adopts the server-minted token — so bridge and row - Network accessline 326
curl http://127.0.0.1:7331/health - Network accessline 371
{ "title": "Docs homepage visual edit", "devServerUrl": "http://localhost:5173", - Network accessline 372
"title": "Docs homepage visual edit", "devServerUrl": "http://localhost:5173", "bridgeUrl": "http://127.0.0.1:7331", - Network accessline 373
"devServerUrl": "http://localhost:5173", - Network accessline 374
"bridgeUrl": "http://127.0.0.1:7331", - Network accessline 394
{ "title": "Tasks responsive visual edit", "devServerUrl": "http://localhost:5173", - Network accessline 395
"title": "Tasks responsive visual edit", "devServerUrl": "http://localhost:5173", "bridgeUrl": "http://127.0.0.1:7331", - Network accessline 396
"devServerUrl": "http://localhost:5173", - Network accessline 397
"bridgeUrl": "http://127.0.0.1:7331", - Network accessline 431
"designId": "<existing-design-id>", "connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:5173", - Network accessline 432
"connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:5173", "paths": ["/settings", "/team"], - Network accessline 433
"devServerUrl": "http://localhost:5173", - Network accessline 447
"designId": "<existing-design-id>", "connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:1234", - Network accessline 448
"connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:1234", "routes": [ - Network accessline 449
"devServerUrl": "http://localhost:1234", - Network accessline 470
The fallback still targets `https://design.agent-native.com`; only the app and
.agents/plugins/agent-native-turn-into-app/.claude-plugin/plugin.json
- Network accessline 5
"author": { "name": "Agent-Native", "url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" - Network accessline 6
"name": "Agent-Native", "url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" }, - Network accessline 7
"url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" - Network accessline 8
}, "homepage": "https://dispatch.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-homepage", "repository": "https://github.com/BuilderIO/agent-native", - Network accessline 9
"homepage": "https://dispatch.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-homepage", - Network accessline 10
"repository": "https://github.com/BuilderIO/agent-native",
.agents/plugins/agent-native-turn-into-app/.mcp.json
.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/attachments.md
.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
.agents/plugins/agent-native-visual-plans/.claude-plugin/plugin.json
- Network accessline 5
"author": { "name": "Agent-Native", "url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" - Network accessline 6
"name": "Agent-Native", "url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" }, - Network accessline 7
"url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" - Network accessline 8
}, "homepage": "https://plan.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-homepage", "repository": "https://github.com/BuilderIO/agent-native", - Network accessline 9
"homepage": "https://plan.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-homepage", - Network accessline 10
"repository": "https://github.com/BuilderIO/agent-native",
.agents/plugins/agent-native-visual-plans/.mcp.json
- Network accessline 1
{ "mcpServers": { "plan": { "type": "http", "url": "https://plan.agent-native.com/mcp" } - Network accessline 2
"mcpServers": { "plan": { "type": "http", "url": "https://plan.agent-native.com/mcp" } } - Network accessline 3
"plan": { "type": "http", "url": "https://plan.agent-native.com/mcp" }
.agents/plugins/agent-native-visual-plans/skills/visual-plan/SKILL.md
- Network accessline 509
separate slash commands. Pass `--no-connect` to register the connector without authenticating, then run `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 510
authenticating, then run `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` whenever you are ready, or choose a narrower `--client`. Auth and MCP tool - Network accessline 511
`npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 535
Hosted default: connect `https://plan.agent-native.com/mcp`. Do
.agents/plugins/agent-native-visual-plans/skills/visual-plan/references/connection.md
.agents/plugins/agent-native-visual-plans/skills/visual-plan/references/local-files.md
- Network accessline 76
is `false`, verify used the offline lint because the app URL was remote or the local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on - Network accessline 77
local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the - Network accessline 78
`--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the concrete validation error.
.agents/plugins/agent-native-visual-plans/skills/visual-recap/SKILL.md
- Network accessline 264
A recap lives only in the database of the MCP that created it. A separately running local dev server (e.g. `http://localhost:8081`) has its OWN database and - Network accessline 265
running local dev server (e.g. `http://localhost:8081`) has its OWN database and - Network accessline 273
`visualUrl` in the returned `plan.mdx` frontmatter, else `url`/`path` resolved against the MCP server's own origin (for the hosted MCP that is `https://plan.agent-native.com`). This always points at the database that has - Network accessline 274
resolved against the MCP server's own origin (for the hosted MCP that is `https://plan.agent-native.com`). This always points at the database that has the plan. - Network accessline 275
`https://plan.agent-native.com`). This always points at the database that has - Network accessline 277
2. Use a `localhost`/dev origin ONLY when the recap was created through a Plan MCP bound to that same origin — i.e. that MCP's url is `http://localhost:<port>/mcp`. Creating through the hosted MCP - Network accessline 278
MCP bound to that same origin — i.e. that MCP's url is `http://localhost:<port>/mcp`. Creating through the hosted MCP and linking to localhost is the exact mismatch that 404s. - Network accessline 279
`http://localhost:<port>/mcp`. Creating through the hosted MCP - Network accessline 280
and linking to localhost is the exact mismatch that 404s. 3. If only a plan id is available, build the MCP origin's absolute URL (hosted: `https://plan.agent-native.com/plans/<id>`) and say it was inferred. - Network accessline 282
(hosted: `https://plan.agent-native.com/plans/<id>`) and say it was inferred. - Network accessline 285
hosted MCP, say so plainly: the local dev server cannot see it. To view a recap on localhost (e.g. to exercise un-deployed local renderer changes), they must connect a LOCAL Plan MCP (`http://localhost:<port>/mcp`) and - Network accessline 286
on localhost (e.g. to exercise un-deployed local renderer changes), they must connect a LOCAL Plan MCP (`http://localhost:<port>/mcp`) and re-create the recap through it so it lands in the local database; offer to do - Network accessline 287
connect a LOCAL Plan MCP (`http://localhost:<port>/mcp`) and
.agents/plugins/agent-native-visual-plans/skills/visual-recap/references/connection.md
.agents/plugins/agent-native-visual-plans/skills/visual-recap/references/local-files.md
- Network accessline 76
is `false`, verify used the offline lint because the app URL was remote or the local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on - Network accessline 77
local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the - Network accessline 78
`--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the concrete validation error.
.agents/plugins/agent-native/.claude-plugin/plugin.json
- Network accessline 5
"author": { "name": "Agent-Native", "url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" - Network accessline 6
"name": "Agent-Native", "url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" }, - Network accessline 7
"url": "https://agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-author" - Network accessline 8
}, "homepage": "https://dispatch.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-homepage", "repository": "https://github.com/BuilderIO/agent-native", - Network accessline 9
"homepage": "https://dispatch.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=plugin-homepage", - Network accessline 10
"repository": "https://github.com/BuilderIO/agent-native",
.agents/plugins/agent-native/.mcp.json
.agents/plugins/agent-native/skills/an/SKILL.md
.claude-plugin/marketplace.json
- Network accessline 10
"source": "./.agents/plugins/agent-native", "autoUpdate": true, "homepage": "https://dispatch.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", - Network accessline 11
"autoUpdate": true, "homepage": "https://dispatch.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", "keywords": [ - Network accessline 12
"homepage": "https://dispatch.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", - Network accessline 25
"source": "./.agents/plugins/agent-native-visual-plans", "autoUpdate": true, "homepage": "https://plan.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", - Network accessline 26
"autoUpdate": true, "homepage": "https://plan.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", "keywords": [ - Network accessline 27
"homepage": "https://plan.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", - Network accessline 40
"source": "./.agents/plugins/agent-native-design", "autoUpdate": true, "homepage": "https://design.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", - Network accessline 41
"autoUpdate": true, "homepage": "https://design.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", "keywords": [ - Network accessline 42
"homepage": "https://design.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", - Network accessline 55
"source": "./.agents/plugins/agent-native-turn-into-app", "autoUpdate": true, "homepage": "https://dispatch.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", - Network accessline 56
"autoUpdate": true, "homepage": "https://dispatch.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage", "keywords": [ - Network accessline 57
"homepage": "https://dispatch.agent-native.com/?utm_source=claude-plugins&utm_medium=referral&utm_content=marketplace-homepage",
community-templates/account-expert/.agents/skills/actions/SKILL.md
community-templates/account-expert/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
community-templates/account-expert/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
community-templates/account-expert/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
community-templates/account-expert/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
community-templates/account-expert/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
community-templates/account-expert/.agents/skills/turn-into-app/references/attachments.md
community-templates/account-expert/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
community-templates/account-expert/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
community-templates/account-expert/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
community-templates/account-expert/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
community-templates/account-expert/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
community-templates/account-tiering/.agents/skills/actions/SKILL.md
community-templates/account-tiering/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
community-templates/account-tiering/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
community-templates/account-tiering/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
community-templates/account-tiering/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
community-templates/account-tiering/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
community-templates/account-tiering/.agents/skills/turn-into-app/references/attachments.md
community-templates/account-tiering/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
community-templates/account-tiering/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
community-templates/account-tiering/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
community-templates/account-tiering/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
community-templates/account-tiering/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
community-templates/call-follow-up-drafter/.agents/skills/actions/SKILL.md
community-templates/call-follow-up-drafter/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
community-templates/call-follow-up-drafter/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
community-templates/call-follow-up-drafter/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
community-templates/call-follow-up-drafter/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
community-templates/call-follow-up-drafter/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
community-templates/call-follow-up-drafter/.agents/skills/turn-into-app/references/attachments.md
community-templates/call-follow-up-drafter/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
community-templates/call-follow-up-drafter/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
community-templates/call-follow-up-drafter/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
community-templates/call-follow-up-drafter/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
community-templates/call-follow-up-drafter/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
community-templates/churn-early-warning/.agents/skills/actions/SKILL.md
community-templates/churn-early-warning/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
community-templates/churn-early-warning/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
community-templates/churn-early-warning/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
community-templates/churn-early-warning/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
community-templates/churn-early-warning/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
community-templates/churn-early-warning/.agents/skills/turn-into-app/references/attachments.md
community-templates/churn-early-warning/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
community-templates/churn-early-warning/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
community-templates/churn-early-warning/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
community-templates/churn-early-warning/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
community-templates/churn-early-warning/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
community-templates/demo-clip-library/.agents/skills/actions/SKILL.md
community-templates/demo-clip-library/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
community-templates/demo-clip-library/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
community-templates/demo-clip-library/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
community-templates/demo-clip-library/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
community-templates/demo-clip-library/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
community-templates/demo-clip-library/.agents/skills/turn-into-app/references/attachments.md
community-templates/demo-clip-library/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
community-templates/demo-clip-library/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
community-templates/demo-clip-library/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
community-templates/demo-clip-library/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
community-templates/demo-clip-library/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
community-templates/linkedin-icp-prospect-tracker/.agents/skills/actions/SKILL.md
community-templates/linkedin-icp-prospect-tracker/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
community-templates/linkedin-icp-prospect-tracker/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
community-templates/linkedin-icp-prospect-tracker/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
community-templates/linkedin-icp-prospect-tracker/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
community-templates/linkedin-icp-prospect-tracker/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
community-templates/linkedin-icp-prospect-tracker/.agents/skills/turn-into-app/references/attachments.md
community-templates/linkedin-icp-prospect-tracker/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
community-templates/linkedin-icp-prospect-tracker/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
community-templates/linkedin-icp-prospect-tracker/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
community-templates/linkedin-icp-prospect-tracker/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
community-templates/linkedin-icp-prospect-tracker/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
community-templates/linkedin-signal-watch/.agents/skills/actions/SKILL.md
community-templates/linkedin-signal-watch/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
community-templates/linkedin-signal-watch/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
community-templates/linkedin-signal-watch/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
community-templates/linkedin-signal-watch/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
community-templates/linkedin-signal-watch/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
community-templates/linkedin-signal-watch/.agents/skills/turn-into-app/references/attachments.md
community-templates/linkedin-signal-watch/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
community-templates/linkedin-signal-watch/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
community-templates/linkedin-signal-watch/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
community-templates/linkedin-signal-watch/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
community-templates/linkedin-signal-watch/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
community-templates/outbound-in-your-voice/.agents/skills/actions/SKILL.md
community-templates/outbound-in-your-voice/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
community-templates/outbound-in-your-voice/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
community-templates/outbound-in-your-voice/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
community-templates/outbound-in-your-voice/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
community-templates/outbound-in-your-voice/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
community-templates/outbound-in-your-voice/.agents/skills/turn-into-app/references/attachments.md
community-templates/outbound-in-your-voice/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
community-templates/outbound-in-your-voice/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
community-templates/outbound-in-your-voice/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
community-templates/outbound-in-your-voice/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
community-templates/outbound-in-your-voice/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
community-templates/win-loss-memo/.agents/skills/actions/SKILL.md
community-templates/win-loss-memo/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
community-templates/win-loss-memo/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
community-templates/win-loss-memo/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
community-templates/win-loss-memo/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
community-templates/win-loss-memo/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
community-templates/win-loss-memo/.agents/skills/turn-into-app/references/attachments.md
community-templates/win-loss-memo/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
community-templates/win-loss-memo/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
community-templates/win-loss-memo/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
community-templates/win-loss-memo/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
community-templates/win-loss-memo/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
packages/core/docs/content/a2a-protocol.mdx
- Access to secretsline 99
[Credential-related source omitted] - Network accessline 123
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 124
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 125
"url": "https://analytics.example.com", - Network accessline 322
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 358
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 359
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 360
"https://analytics.example.com", - Access to secretsline 362
[Credential-related source omitted] - Network accessline 380
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", - Network accessline 381
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", "cardUrl": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a/agentCard/v1.0", - Network accessline 382
"url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", - Network accessline 383
"cardUrl": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a/agentCard/v1.0", - Network accessline 420
[Azure AI Foundry hosted agents](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 426
Foundry can expose an A2A endpoint through [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint). - Network accessline 436
For service-principal client credentials, request the Entra scope `https://ai.azure.com/.default` and grant the caller the **Foundry Agent - Network accessline 437
`https://ai.azure.com/.default` and grant the caller the **Foundry Agent - Network accessline 443
"auth": { "type": "oauth-client-credentials", "tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", - Network accessline 444
"type": "oauth-client-credentials", "tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", "clientId": "<application-id>", - Network accessline 445
"tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", - Network accessline 446
"clientId": "<application-id>", "clientSecretRef": "FOUNDRY_CLIENT_SECRET", "scope": "https://ai.azure.com/.default" - Network accessline 447
[Credential-related source omitted] - Access to secretsline 447
[Credential-related source omitted] - Network accessline 448
"scope": "https://ai.azure.com/.default" - Network accessline 455
[Gemini Enterprise can register external A2A agents](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent), - Network accessline 468
permission. Custom A2A-agent registration is Pre-GA, so confirm availability in the target project before exposing it in a workspace picker. See [invoke a Gemini Enterprise agent](https://docs.cloud.google.com/gemini/enterprise/docs/invoke- - Network accessline 469
the target project before exposing it in a workspace picker. See [invoke a Gemini Enterprise agent](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) for the resource URL and IAM requirements. - Network accessline 470
Gemini Enterprise agent](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) - Network accessline 486
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 487
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 488
"url": "https://api.anthropic.com", - Access to secretsline 493
[Credential-related source omitted] - Access to secretsline 511
[Credential-related source omitted] - Network accessline 557
# Call a deployed app directly by URL, reading the bearer token from env agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 558
[Credential-related source omitted] - Access to secretsline 558
[Credential-related source omitted] - Access to secretsline 559
[Credential-related source omitted] - Access to secretsline 631
[Credential-related source omitted] - Access to secretsline 635
[Credential-related source omitted] - Network accessline 693
- A shared `A2A_SECRET` is mounted into every app's environment at build time. - Cross-app calls are same-origin — `https://workspace.example.com/apps/analytics` calls `https://workspace.example.com/apps/mail` — so there is no DNS, CORS, or - Network accessline 694
- Cross-app calls are same-origin — `https://workspace.example.com/apps/analytics` calls `https://workspace.example.com/apps/mail` — so there is no DNS, CORS, or per-pair JWT setup. - Network accessline 702
**Never rely on a fire-and-forget `Promise` outliving the response.** Serverless functions (Netlify, Vercel, AWS Lambda, Cloud Run) freeze the moment the response body is flushed — sometimes before the TCP handshake of an unawaited `fetch(. - Network accessline 740
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 741
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 742
"https://analytics.example.com", - Access to secretsline 744
[Credential-related source omitted]
packages/core/docs/content/actions-defining.mdx
- Network accessline 226
`schema` accepts any [Standard Schema](https://standardschema.dev)-compatible library:
packages/core/docs/content/add-ui-to-your-agent.mdx
- Network accessline 27
Foundry exposes an A2A Agent Card at the agent endpoint. Use the v1.0 JSON-RPC card URL, select **A2A**, then choose either a bearer credential or OAuth client credentials. For service-to-service access, request the `https://ai.azure.com/.d - Network accessline 45
The directory links to the [Global A2A Registry](https://www.a2a-registry.org), a public community catalog. Registry entries are discovery only. Review the Agent Card, verify the host and auth requirements, and explicitly add the agent befo
packages/core/docs/content/agent-mentions.mdx
packages/core/docs/content/agent-native-config.mdx
- Access to secretsline 186
[Credential-related source omitted] - Access to secretsline 187
[Credential-related source omitted] - Access to secretsline 352
[Credential-related source omitted] - Network accessline 454
export default defineAgentNativeConfig({ runtime: { apiBaseUrl: "https://api.example.com", - Network accessline 455
runtime: { apiBaseUrl: "https://api.example.com", }, - Network accessline 456
apiBaseUrl: "https://api.example.com", - Network accessline 461
The value is an absolute `http` or `https` URL with an optional mount path and no query string, fragment, or credentials; an empty string keeps same-origin requests. It is written into the browser bundle at build time, or a static host can - Network accessline 463
With a base URL set, `agentNativePath()` and `appApiPath()` return absolute URLs on the server, so client code must build every server URL — `fetch`, `XMLHttpRequest`, `EventSource`, image and download links — with those helpers. A hand-wri
packages/core/docs/content/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/agent-web-surfaces.mdx
- Network accessline 171
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 172
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 173
siteUrl: "https://example.com", - Network accessline 198
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/agentkit.mdx
- Network accessline 18
client, a custom transport, capability-gated UI, or the HTTP contract for a non-Agent-Native backend. For exhaustive API reference, see the package [README](https://github.com/BuilderIO/agent-native/blob/main/packages/agentkit/README.md) - Network accessline 19
non-Agent-Native backend. For exhaustive API reference, see the package [README](https://github.com/BuilderIO/agent-native/blob/main/packages/agentkit/README.md) and [ARCHITECTURE.md](https://github.com/BuilderIO/agent-native/blob/main/pack - Network accessline 20
[README](https://github.com/BuilderIO/agent-native/blob/main/packages/agentkit/README.md) - Network accessline 21
and [ARCHITECTURE.md](https://github.com/BuilderIO/agent-native/blob/main/packages/agentkit/ARCHITECTURE.md). - Network accessline 275
- [**Package README**](https://github.com/BuilderIO/agent-native/blob/main/packages/agentkit/README.md) — exhaustive reference: connection requests, approval decisions, durable snapshots and replay, slash discovery, semantic styling, and mi - Network accessline 276
- [**ARCHITECTURE.md**](https://github.com/BuilderIO/agent-native/blob/main/packages/agentkit/ARCHITECTURE.md) — ownership, invariants, and the extension model.
packages/core/docs/content/audit-log.mdx
- Network accessline 211
**`export-audit-ocsf`** returns the organization's trail as [OCSF](https://schema.ocsf.io) 1.9.0 API Activity events (`class_uid` 6003), oldest first, for a SIEM to pull on a schedule. Owners and admins only, and it reads the same `org` and
packages/core/docs/content/authentication.mdx
- Network accessline 10
Agent-Native apps use [Better Auth](https://better-auth.com) for authentication with an account-first design. Users create an account on first visit and get real identity from day one. - Access to secretsline 251
[Credential-related source omitted] - Access to secretsline 256
[Credential-related source omitted] - Access to secretsline 260
[Credential-related source omitted] - Access to secretsline 265
[Credential-related source omitted] - Access to secretsline 267
[Credential-related source omitted] - Access to secretsline 289
[Credential-related source omitted] - Access to secretsline 293
[Credential-related source omitted] - Network accessline 306
https://mail.agent-native.com/mcp - Access to secretsline 398
[Credential-related source omitted] - Access to secretsline 531
[Credential-related source omitted] - Access to secretsline 533
[Credential-related source omitted] - Access to secretsline 535
[Credential-related source omitted] - Access to secretsline 536
[Credential-related source omitted]
packages/core/docs/content/automation-connectors.mdx
- Network accessline 55
}, outbound: { baseUrl: "https://n8n.example.com", - Network accessline 56
outbound: { baseUrl: "https://n8n.example.com", path: "/webhook/notify-sales", - Network accessline 57
baseUrl: "https://n8n.example.com", - Network accessline 81
The agent calls this action with `{ workflowId: "notify-sales-slack", input: { leadEmail: "jane@acme.com" } }` — it never sees `https://n8n.example.com/webhook/notify-sales` or the bearer token. `outbound.baseUrl` + `outbound.path` are stat - Network accessline 108
See the [n8n Webhook documentation](https://docs.n8n.io/integrations/builtin/core-nodes/n8n-nodes-base.webhook/). - Network accessline 118
See [Zapier REST Hook triggers](https://docs.zapier.com/integrations/build/hook-trigger) and [Zapier MCP](https://docs.zapier.com/mcp/home).
packages/core/docs/content/automations.mdx
packages/core/docs/content/aws-amplify.mdx
- Access to secretsline 116
[Credential-related source omitted] - Network accessline 133
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 195
[Credential-related source omitted] - Network accessline 209
Create the app in the [Amplify console](https://console.aws.amazon.com/amplify/): - Network accessline 244
The first build takes a few minutes, and you can follow it in the Amplify console. When it finishes, open the app's address, which looks like `https://main.d1m7bkiki6tdw1.amplifyapp.com`, and create the first account. - Network accessline 246
`https://main.d1m7bkiki6tdw1.amplifyapp.com`, and create the first account. - Access to secretsline 298
[Credential-related source omitted] - Access to secretsline 302
[Credential-related source omitted]
packages/core/docs/content/aws-lambda.mdx
- Network accessline 39
You also need an AWS account and the [AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html) - Access to secretsline 117
[Credential-related source omitted] - Access to secretsline 152
[Credential-related source omitted] - Access to secretsline 170
[Credential-related source omitted] - Access to secretsline 218
[Credential-related source omitted] - Access to secretsline 226
[Credential-related source omitted] - Network accessline 251
The first command prints the `FunctionUrl`, which looks like `https://abc123.lambda-url.us-east-1.on.aws/`. Copy it for Step 7. The two `add-permission` commands allow requests to reach the function through that - Network accessline 252
`https://abc123.lambda-url.us-east-1.on.aws/`. Copy it for Step 7. The two `add-permission` commands allow requests to reach the function through that address. - Network accessline 271
Create the CloudFront distribution in the [CloudFront console](https://console.aws.amazon.com/cloudfront/v4/home): - Network accessline 272
[CloudFront console](https://console.aws.amazon.com/cloudfront/v4/home): - Network accessline 286
On the distribution's **Origins** tab, create an origin. For **Origin domain**, paste the function URL's host name from Step 5, without `https://` or the - Network accessline 287
paste the function URL's host name from Step 5, without `https://` or the - Access to secretsline 313
[Credential-related source omitted] - Network accessline 323
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 325
[Credential-related source omitted] - Access to secretsline 327
[Credential-related source omitted] - Access to secretsline 344
[Credential-related source omitted] - Dangerous shell operationline 346
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/aws-rds.mdx
- Network accessline 40
- [Database](/docs/server-database) - schema, migrations, and connection rules - [Deployment](/docs/deployment) - choose a runtime and deployment preset - [Amazon RDS for PostgreSQL connection docs](https://docs.aws.amazon.com/AmazonRDS/lat - Network accessline 42
- [Amazon RDS for PostgreSQL connection docs](https://docs.aws.amazon.com/AmazonRDS/latest/UserGuide/USER_ConnectToPostgreSQLInstance.html) - find the endpoint and connect securely
packages/core/docs/content/azure-postgresql.mdx
- Network accessline 41
- [Database](/docs/server-database) - schema, migrations, and connection rules - [Deployment](/docs/deployment) - choose a runtime and deployment preset - [Azure Database for PostgreSQL connection docs](https://learn.microsoft.com/en-us/azu - Network accessline 43
- [Azure Database for PostgreSQL connection docs](https://learn.microsoft.com/en-us/azure/postgresql/overview) - configure the server and networking
packages/core/docs/content/azure-static-web-apps.mdx
packages/core/docs/content/blueprint-installer.mdx
- Network accessline 72
agent-native add <kind> <name> # print a curated blueprint agent-native add <kind> <https://docs…> # research-and-integrate from a URL - Network accessline 73
agent-native add <kind> <https://docs…> # research-and-integrate from a URL - Network accessline 99
agent-native add provider https://docs.example.com/api | claude
packages/core/docs/content/cli-adapters.mdx
- Access to secretsline 103
[Credential-related source omitted]
packages/core/docs/content/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">Component</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 126
optional: true, description: "Everything from the [useQuery options](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) except `queryKey` and `queryFn`, which the hook sets itself.", - Network accessline 127
description: "Everything from the [useQuery options](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) except `queryKey` and `queryFn`, which the hook sets itself.", }, - Network accessline 128
"Everything from the [useQuery options](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) except `queryKey` and `queryFn`, which the hook sets itself.", - Network accessline 138
This hook is intended to be used for read-only actions. It is a thin wrapper around the [useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) function. - Network accessline 174
optional: true, description: "Everything from the [useMutation options](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) except `mutationFn`, which the hook sets itself.", - Network accessline 175
description: "Everything from the [useMutation options](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) except `mutationFn`, which the hook sets itself.", fields: [ - Network accessline 176
"Everything from the [useMutation options](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) except `mutationFn`, which the hook sets itself.", - Network accessline 202
This hook is intended to be used for actions that change data, like creating, updating, or sending something. It is a thin wrapper around the [useMutation()](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) func
packages/core/docs/content/cloneable-saas.mdx
- Network accessline 48
https://github.com/owner/repo - Network accessline 65
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 66
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 74
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 75
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 146
Want to be listed? [Submit a community template](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) - Access to secretsline 307
[Credential-related source omitted]
packages/core/docs/content/cloud-sql.mdx
- Network accessline 23
If the deployment runs the [Cloud SQL Auth Proxy](https://cloud.google.com/sql/docs/postgres/connect-auth-proxy) - Network accessline 53
- [Database](/docs/server-database) - schema, migrations, and connection rules - [Deployment](/docs/deployment) - choose a runtime and deployment preset - [Cloud SQL for PostgreSQL connection options](https://cloud.google.com/sql/docs/postg - Network accessline 55
- [Cloud SQL for PostgreSQL connection options](https://cloud.google.com/sql/docs/postgres/connect-overview) - choose direct, proxy, or connector access
packages/core/docs/content/cloudflare.mdx
- Access to secretsline 101
[Credential-related source omitted] - Network accessline 118
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 168
[Credential-related source omitted] - Network accessline 211
Scheduled automations, queued event automations, and cleanup of interrupted runs all happen on a once-a-minute sweep. The build adds a `* * * * *` [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) - Network accessline 212
runs all happen on a once-a-minute sweep. The build adds a `* * * * *` [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) to the generated `wrangler.json` and keeps any Cron Triggers you already - Network accessline 213
[Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/)
packages/core/docs/content/code-agents-ui.mdx
- Network accessline 205
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 206
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 369
<div class="diagram-col"> <div class="diagram-pill">./local-app</div> <div class="diagram-pill">https://example.com</div> - Network accessline 370
<div class="diagram-pill">./local-app</div> <div class="diagram-pill">https://example.com</div> <div class="diagram-pill">--describe \"...\"</div> - Network accessline 371
<div class="diagram-pill">https://example.com</div> - Network accessline 414
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 415
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard"
packages/core/docs/content/create-with-an-agent.mdx
- Network accessline 14
- An AI agent that can run terminal commands, such as Claude Code, Codex, or Cursor. Agents that only chat, such as ChatGPT or Claude on the web, can still help. See [Agents that can't run commands](#chat-only-agents). - [Node.js](https://n - Network accessline 15
- [Node.js](https://nodejs.org) 22.22 or later. - Network accessline 37
5. Start the app with `pnpm dev` and keep it running in the background. 6. Read the local URL that `pnpm dev` prints. Check that it responds, for example with `curl -I <url>`. Do not open a browser and do not - Network accessline 38
6. Read the local URL that `pnpm dev` prints. Check that it responds, for example with `curl -I <url>`. Do not open a browser and do not use headless browser tools. I will open the app myself. - Network accessline 39
for example with `curl -I <url>`. Do not open a browser and do not - Access to secretsline 48
[Credential-related source omitted] - Network accessline 57
- A new app folder, named `my-app` unless you changed it, that contains `actions/`, `app/`, and an `AGENTS.md` file. - `pnpm dev` running without errors. - A local URL, usually `http://localhost:8080`, that the agent has checked responds. - Network accessline 59
- A local URL, usually `http://localhost:8080`, that the agent has checked responds.
packages/core/docs/content/creating-templates.mdx
packages/core/docs/content/cross-app-connect-design.mdx
- Network accessline 44
When the framework flag `labs.connectApps` is enabled, Dispatch exposes a **Connect apps** page at `/connect`. It reads the public catalog from `https://www.agent-native.com/apps.json`, caches the feed, and lets a user - Network accessline 45
**Connect apps** page at `/connect`. It reads the public catalog from `https://www.agent-native.com/apps.json`, caches the feed, and lets a user inspect an app's `/.well-known/agent-card.json` before starting the existing - Network accessline 46
`https://www.agent-native.com/apps.json`, caches the feed, and lets a user - Access to secretsline 63
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted]
packages/core/docs/content/cross-app-sso.mdx
packages/core/docs/content/decisions.mdx
- Network accessline 16
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 17
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 18
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/deno-deploy.mdx
- Network accessline 27
You also need a Deno Deploy account and [Deno](https://docs.deno.com/runtime/getting_started/installation/) - Access to secretsline 103
[Credential-related source omitted] - Network accessline 120
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 171
[Credential-related source omitted]
packages/core/docs/content/deployment-environment-variables.mdx
- Network accessline 31
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | Skip on-demand schema DDL on deployments the automatic serverless skip does not cover. Requires that your schema is already migrated. See [On-demand schema DDL](/docs/deployment#skip-ensure-ta - Network accessline 33
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | Corporate callback-only, comma-separated allowlist of exact trusted preview origins (for example, `https://0123456789abcdef01234567--content.netlify.app`). Set it only on the corporate callbac - Network accessline 63
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 64
| `BETTER_AUTH_SECRET` | 32+ char random string. Signs session cookies and is the fallback for `OAUTH_STATE_SECRET` and app-local encrypted values. It is also a legacy read candidate for existing shared vault cipherte - Network accessline 65
| `BETTER_AUTH_URL` | Optional public origin override for this app (e.g. `https://mail.example.com`). The framework infers it from request, known-template, and hosting-platform context when unset. - Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 67
[Credential-related source omitted] - Access to secretsline 75
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Access to secretsline 97
[Credential-related source omitted]
packages/core/docs/content/deployment.mdx
packages/core/docs/content/dispatch.mdx
- Network accessline 120
Dispatch can be the single MCP connector for external agents: add `https://dispatch.agent-native.com/mcp` once in Claude, ChatGPT, Codex, or Cursor, and one authorization reaches every granted workspace app instead of one connector per app. - Network accessline 129
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p
packages/core/docs/content/docker.mdx
- Dangerous shell operationline 73
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 75
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 116
[Credential-related source omitted] - Access to secretsline 172
[Credential-related source omitted] - Network accessline 189
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/docs-components.mdx
- Network accessline 645
Use a fenced code block with the `mermaid` language tag. Mermaid supports flowcharts (`graph`), sequence diagrams (`sequenceDiagram`), entity-relationship diagrams (`erDiagram`), and more. See [mermaid.js.org](https://mermaid.js.org/intro/) - Network accessline 809
id: "manual-code", data: { code: "const res = await fetch('/api/analyses')\nconst data = await res.json()", - Network accessline 810
data: { code: "const res = await fetch('/api/analyses')\nconst data = await res.json()", language: "ts", - Network accessline 811
code: "const res = await fetch('/api/analyses')\nconst data = await res.json()", - Network accessline 844
id: "manual-code", data: { code: "const res = await fetch('/api/analyses')\nconst data = await res.json()", - Network accessline 845
data: { code: "const res = await fetch('/api/analyses')\nconst data = await res.json()", language: "ts", - Network accessline 846
code: "const res = await fetch('/api/analyses')\nconst data = await res.json()",
packages/core/docs/content/doctor.mdx
packages/core/docs/content/drafts/clips-experimental-features.md
packages/core/docs/content/durable-background-runs.mdx
- Network accessline 168
URL, then `APP_URL`), which is a round trip through the host's edge. Set `AGENT_NATIVE_SELF_DISPATCH_URL` to the address this process listens on, such as `http://127.0.0.1:3000`, to dispatch over loopback instead. It must be an - Network accessline 169
`AGENT_NATIVE_SELF_DISPATCH_URL` to the address this process listens on, such as `http://127.0.0.1:3000`, to dispatch over loopback instead. It must be an absolute `http(s)` URL, and the app base path is appended to it. It only - Network accessline 170
as `http://127.0.0.1:3000`, to dispatch over loopback instead. It must be an absolute `http(s)` URL, and the app base path is appended to it. It only addresses requests the deployment sends to itself, never a URL an outside - Network accessline 245
request, and Legacy Free and Starter don't include background functions; see Netlify's [functions usage and billing](https://docs.netlify.com/build/functions/usage-and-billing/) - Network accessline 246
Netlify's [functions usage and billing](https://docs.netlify.com/build/functions/usage-and-billing/) page. - Network accessline 247
[functions usage and billing](https://docs.netlify.com/build/functions/usage-and-billing/)
packages/core/docs/content/embedding-sdk.mdx
- Network accessline 221
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 222
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 223
url="https://assets.agent-native.com/picker" - Network accessline 276
getMcpUrl("https://assets.agent-native.com"); - Network accessline 277
getA2AUrl("https://assets.agent-native.com"); - Network accessline 279
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 280
"https://assets.agent-native.com", - Network accessline 305
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 306
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 307
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 389
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 391
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 393
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 394
hostOrigin: "https://app.example.com", - Network accessline 397
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 398
hostOrigin: "https://app.example.com", - Network accessline 402
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 403
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 404
{ hostOrigin: "https://app.example.com" }, - Network accessline 408
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 409
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 410
{ hostOrigin: "https://app.example.com" }, - Network accessline 413
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 414
hostOrigin: "https://app.example.com", - Network accessline 467
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 468
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 469
hostOrigin: "https://app.example.com",
packages/core/docs/content/environment-variables.mdx
- Access to secretsline 4
[Credential-related source omitted] - Access to secretsline 79
[Credential-related source omitted] - Access to secretsline 84
[Credential-related source omitted] - Access to secretsline 95
[Credential-related source omitted] - Network accessline 96
| `AGENT_PROD_CODE_EXECUTION` | Production code-execution policy. | `off`, `sand - Network accessline 97
| `AGENT_NATIVE_SSR_CACHE` | Deployment-wide public SSR shell cache policy. | Unset/`on`, - Network accessline 98
| `AGENT_NATIVE_HEAD_SCRIPT_SRC` | Extra script added to every server-rendered page head. Where the consent banner shows, it waits for analytics consent, like Google Analytics. | An absolute - Network accessline 107
| `COOKIE_DOMAIN` | Shares auth cookies across same-database subdomains when explicitly configured. | A cookie dom - Network accessline 108
| `CORS_ALLOWED_ORIGINS` | Cross-origin request allowlist. Listed origins may send credentials, which a UI on another origin needs (see `runtime.apiBaseUrl`). | A comma-sepa - Network accessline 109
| `AGENT_NATIVE_SELF_DISPATCH_URL` | Where this deployment sends its own background work, ahead of the platform URLs and `APP_URL`. Lets a long-lived server dispatch over loopback instead of its edge. | An absolute - Access to secretsline 159
[Credential-related source omitted] - Access to secretsline 161
[Credential-related source omitted] - Access to secretsline 168
[Credential-related source omitted] - Access to secretsline 192
[Credential-related source omitted] - Access to secretsline 196
[Credential-related source omitted] - Access to secretsline 197
[Credential-related source omitted] - Access to secretsline 198
[Credential-related source omitted] - Access to secretsline 199
[Credential-related source omitted] - Access to secretsline 200
[Credential-related source omitted] - Network accessline 202
Additional provider variables are listed in the [full environment-variable inventory on GitHub](https://github.com/BuilderIO/agent-native/blob/main/docs/environment-variables.md#authentication-provider-and-integration-variables). - Access to secretsline 211
[Credential-related source omitted] - Access to secretsline 212
[Credential-related source omitted] - Access to secretsline 213
[Credential-related source omitted] - Access to secretsline 214
[Credential-related source omitted] - Access to secretsline 215
[Credential-related source omitted] - Access to secretsline 216
[Credential-related source omitted]
packages/core/docs/content/extensions.mdx
- Network accessline 275
| `dbQuery(sql, args)` | Read from SQL (auto-scoped to the user; never for writes) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` | L - Network accessline 276
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` | H - Network accessline 277
| `extensionFetch(url, options)` | Hit external APIs through a secure proxy with secrets | `extensionFetch('https://api.github.com/user')` | - Network accessline 494
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 495
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/external-agents-catalog.mdx
- Network accessline 172
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex --full-catalog - Access to secretsline 471
[Credential-related source omitted] - Access to secretsline 487
[Credential-related source omitted] - Access to secretsline 493
[Credential-related source omitted] - Network accessline 499
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 500
[Credential-related source omitted] - Network accessline 501
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 502
[Credential-related source omitted] - Network accessline 524
`connect dev` rewrites the same stable MCP server names (`agent-native-mail`, `agent-native-calendar`, etc.) to the local dev-lazy gateway, so tool names do not change. It backs up the current production entries in `~/.agent-native/connect- - Access to secretsline 538
[Credential-related source omitted] - Network accessline 551
- Connect your own agent to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; use a direct app URL only when you want one isolated app. - Access to secretsline 571
[Credential-related source omitted]
packages/core/docs/content/external-agents.mdx
- Network accessline 114
https://dispatch.agent-native.com/mcp - Network accessline 125
https://mail.agent-native.com/mcp - Network accessline 129
Every hosted app also has a helper page at `https://<app>/mcp/connect` with the copyable URL and - Network accessline 130
`https://<app>/mcp/connect` with the copyable URL and - Network accessline 256
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 280
When troubleshooting local MCP config, redact `Authorization`, `http_headers`, and token values before sharing logs. Do not use raw curl as a substitute for a - Network accessline 281
and token values before sharing logs. Do not use raw curl as a substitute for a - Network accessline 300
The raw `skills` CLI installs `SKILL.md` files only; local MCP hosts still need a connector such as `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 301
need a connector such as `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 312
The default client selection is all supported local clients; add `--client codex`, `--client claude-code`, or another specific target to narrow setup. Inline hosts (ChatGPT, Claude chat on web, desktop, and mobile, plus the Claude Desktop C - Network accessline 318
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 333
npx @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 336
`reconnect` finds any MCP config entry whose URL ends in `/mcp` for the given host and selected client (matching by URL regardless of connector name), then refreshes or replaces the auth material without touching your installed skills or re - Network accessline 344
For MCP hosts that cannot add a remote OAuth URL directly, open the app in your browser and use its **Connect** affordance (served at `https://<app>/mcp/connect`). While logged in, click **Connect / Authorize**. The page hands you eit - Network accessline 350
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 351
[Credential-related source omitted] - Network accessline 352
"url": "https://mail.agent-native.com/mcp", - Network accessline 368
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 369
https://dispatch.agent-native.com/mcp - Network accessline 372
This is the same URL-only entry that `npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code` writes for you. Then run `/mcp` in Claude Code and choose **Authenticate**. The client discovers auth from t - Network accessline 375
https://dispatch.agent-native.com/mcp - Network accessline 403
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 404
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=…
packages/core/docs/content/file-uploads.mdx
- Network accessline 142
Register a custom provider only for a different kind of backend. This example uploads through a generic HTTP API using `fetch()`: - Network accessline 143
example uploads through a generic HTTP API using `fetch()`: - Network accessline 160
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 161
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 162
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/getting-started.mdx
packages/core/docs/content/harness-agents.mdx
- Network accessline 120
Agent-Native can act as an [ACP](https://agentclientprotocol.com) (Agent Client
packages/core/docs/content/http-api.mdx
- Network accessline 2
title: "HTTP API: Call Actions from Anything" description: "Call an agent-native app's actions over plain HTTP with a long-lived bearer token — from a support-desk bot, a CI job, a cron script, or any backend service, no browser session req - Network accessline 3
description: "Call an agent-native app's actions over plain HTTP with a long-lived bearer token — from a support-desk bot, a CI job, a cron script, or any backend service, no browser session required." search: "HTTP API bearer token service - Network accessline 4
search: "HTTP API bearer token service token connect CLI curl action endpoint server-to-server webhook CI backend integration Authorization header revoke rotate" - Network accessline 41
npx @agent-native/core@latest connect https://clips.example.com - Network accessline 49
npx @agent-native/core@latest connect https://clips.example.com \ - Access to secretsline 65
[Credential-related source omitted] - Network accessline 74
[Credential-related source omitted] - Access to secretsline 74
[Credential-related source omitted] - Network accessline 75
https://clips.example.com/_agent-native/actions/import-loom-recording \ - Network accessline 76
[Credential-related source omitted] - Access to secretsline 76
[Credential-related source omitted] - Network accessline 78
-d '{"url": "https://www.loom.com/share/EXAMPLE_SHARE_ID"}' - Access to secretsline 89
[Credential-related source omitted] - Access to secretsline 95
[Credential-related source omitted] - Network accessline 106
"example": "{\n \"url\": \"https://www.loom.com/share/EXAMPLE_SHARE_ID\"\n}" - Access to secretsline 134
[Credential-related source omitted] - Network accessline 154
[Credential-related source omitted] - Access to secretsline 154
[Credential-related source omitted] - Network accessline 155
"https://crm.example.com/_agent-native/actions/get-lead?leadId=lead_123" \ - Access to secretsline 156
[Credential-related source omitted] - Network accessline 189
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 198
const res = await fetch( - Network accessline 199
"https://clips.example.com/_agent-native/actions/import-loom-recording", - Access to secretsline 203
[Credential-related source omitted] - Network accessline 218
**3. Link back to the recording.** Post `https://clips.example.com/r/${recordingId}` into the ticket thread. Because the bot used an org service token, the recording is org-scoped: every teammate who opens that link sees it, and it does not - Network accessline 273
| `403` on a data action from a service token only | The app authorizes with its own `org_members` role lookup and has not adopted `implicitServiceOrgRole` from `@agent-native/core/org`, so the synthetic identity resolves no - Network accessline 274
| `403` on `create-org-service-token` / `revoke-org-service-token` | The caller is an org `member`, or is a service token. Minting and revoking require a human owner/admin. - Network accessline 275
| `403` from browser JavaScript, but curl works | Cross-origin browser call from an origin that is not allowlisted. See [CSRF and CORS](#csrf-cors). - Access to secretsline 282
[Credential-related source omitted]
packages/core/docs/content/integrations.mdx
- Network accessline 335
| Integration | Connection mode | Status | Endpoi - Network accessline 336
| ----------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ---------------------- | ---------------------------------- | ------ - Network accessline 337
| [Context7](https://context7.com/) | No auth | Ready · verified | `https - Network accessline 338
| [Sentry](https://docs.sentry.io/product/sentry-mcp/) | Header / token | Ready · preflight only | `https - Network accessline 339
| [Notion](https://developers.notion.com/guides/mcp/get-started-with-mcp) | OAuth | Ready · preflight only | `https - Network accessline 340
| [Granola](https://docs.granola.ai/help-center/sharing/integrations/mcp) | OAuth | Ready · preflight only | `https - Network accessline 341
| [Semgrep](https://github.com/semgrep/mcp#readme) | No auth | Ready · preflight only | `https - Network accessline 342
| [Linear](https://linear.app/docs/mcp) | OAuth | Ready · preflight only | `https - Network accessline 343
| [Atlassian Rovo MCP](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) | OAuth + provider setup | Restricted | `https - Network accessline 344
| [Supabase](https://www.builder.io/c/docs/fusion-connect-to-supabase?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=fusion_connect_supabase) | OAuth | Ready · preflight only | `https - Network accessline 345
| [Neon](https://www.builder.io/c/docs/fusion-connect-to-neon?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=fusion_connect_neon) | OAuth | Ready · preflight only | `https - Network accessline 346
| [Stripe](https://docs.stripe.com/mcp) | OAuth | Ready · preflight only | `https - Network accessline 347
| [Cloudflare](https://developers.cloudflare.com/agents/model-context-protocol/cloudflare/servers-for-cloudflare/) | OAuth | Ready · preflight only | `https - Network accessline 348
| [GitLab](https://docs.gitlab.com/user/model_context_protocol/mcp_server/) | OAuth + provider setup | Beta · restricted | `https - Network accessline 349
| [Figma](https://developers.figma.com/docs/figma-mcp-server/) | Manual setup | Approved clients only · restricted | `https - Network accessline 350
| [Canva](https://www.canva.dev/docs/mcp/) | Manual setup | Approved domains · restricted | `https - Network accessline 351
| [Vercel](https://vercel.com/docs/agent-resources/vercel-mcp) | Manual setup | Approved clients only · restricted | `https - Network accessline 352
[Credential-related source omitted] - Network accessline 353
| [Slack](https://docs.slack.dev/ai/slack-mcp-server/) | Manual setup | Approved clients only · restricted | `https - Network accessline 354
| [Asana](https://developers.asana.com/docs/integrating-with-asanas-mcp-server) | Manual setup | Provider setup · restricted | `https - Network accessline 355
| [HubSpot](https://developers.hubspot.com/docs/apps/developer-platform/build-apps/integrate-with-the-remote-hubspot-mcp-server) | Manual setup | Provider setup · restricted | `https - Network accessline 356
| [Intercom](https://developers.intercom.com/docs/guides/mcp) | OAuth | Ready · preflight only | `https - Network accessline 357
| [monday.com](https://developer.monday.com/api-reference/docs/build-on-monday-with-ai) | OAuth | Ready · preflight only | `https - Network accessline 358
| [Webflow](https://developers.webflow.com/mcp/reference/getting-started) | OAuth | Ready · preflight only | `https - Network accessline 359
| [PayPal](https://developer.paypal.com/ai-tools/mcp-server/) | OAuth | Ready · preflight only | `https - Network accessline 360
| [Box](https://developer.box.com/guides/box-mcp) | Manual setup | Provider setup · restricted | `https - Network accessline 361
| [Netlify](https://docs.netlify.com/build/build-with-ai/agent-setup-guides/agent-setup-overview/) | OAuth | Ready · preflight only | `https - Network accessline 362
| [Zapier MCP](https://help.zapier.com/hc/en-us/articles/36265392843917-Use-Zapier-MCP-with-your-client) | Header / token | User-created server · restricted | `https
packages/core/docs/content/key-concepts.mdx
- Network accessline 341
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 342
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 343
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/koyeb.mdx
packages/core/docs/content/locales/ar-SA/a2a-protocol.mdx
- Access to secretsline 93
[Credential-related source omitted] - Network accessline 117
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 118
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 119
"url": "https://analytics.example.com", - Network accessline 314
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 350
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 351
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 352
"https://analytics.example.com", - Access to secretsline 354
[Credential-related source omitted] - Network accessline 372
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", - Network accessline 373
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", "cardUrl": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a/agentCard/v1.0", - Network accessline 374
"url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", - Network accessline 375
"cardUrl": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a/agentCard/v1.0", - Network accessline 411
تشغّل [الوكلاء المستضافون في Azure AI Foundry](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 416
يمكن لـ Foundry كشف نقطة نهاية A2A من خلال [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint). - Network accessline 425
لمصادقة بيانات اعتماد كيان الخدمة، اطلب رمز Entra بالنطاق `https://ai.azure.com/.default` وامنح المتصل دور **Foundry Agent Consumer** على - Network accessline 426
`https://ai.azure.com/.default` وامنح المتصل دور **Foundry Agent Consumer** على - Network accessline 431
"auth": { "type": "oauth-client-credentials", "tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", - Network accessline 432
"type": "oauth-client-credentials", "tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", "clientId": "<application-id>", - Network accessline 433
"tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", - Network accessline 434
"clientId": "<application-id>", "clientSecretRef": "FOUNDRY_CLIENT_SECRET", "scope": "https://ai.azure.com/.default" - Network accessline 435
[Credential-related source omitted] - Access to secretsline 435
[Credential-related source omitted] - Network accessline 436
"scope": "https://ai.azure.com/.default" - Network accessline 443
يمكن لـ [Gemini Enterprise تسجيل وكلاء A2A خارجيين](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent)، - Network accessline 455
يحتاج المتصل إلى رمز حامل Google OAuth مع إذن `discoveryengine.assist`. تسجيل وكلاء A2A المخصصين في مرحلة ما قبل التوافر العام، لذا تحقّق من توفره في المشروع المستهدف قبل عرضه في منتقي مساحة العمل. راجع [استدعاء وكيل Gemini Enterprise](http - Network accessline 456
وكلاء A2A المخصصين في مرحلة ما قبل التوافر العام، لذا تحقّق من توفره في المشروع المستهدف قبل عرضه في منتقي مساحة العمل. راجع [استدعاء وكيل Gemini Enterprise](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) لعنوان المو - Network accessline 457
المستهدف قبل عرضه في منتقي مساحة العمل. راجع [استدعاء وكيل Gemini Enterprise](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) - Network accessline 472
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 473
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 474
"url": "https://api.anthropic.com", - Access to secretsline 479
[Credential-related source omitted] - Access to secretsline 497
[Credential-related source omitted] - Network accessline 543
# استدعِ تطبيقًا منشورًا مباشرةً عبر عنوان URL، مع قراءة رمز الحامل من متغير بيئة agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 544
[Credential-related source omitted] - Access to secretsline 544
[Credential-related source omitted] - Access to secretsline 545
[Credential-related source omitted] - Access to secretsline 616
[Credential-related source omitted] - Access to secretsline 620
[Credential-related source omitted] - Network accessline 678
- يتم تثبيت `A2A_SECRET` مشترك في بيئة كل تطبيق في وقت الإنشاء. - المكالمات عبر التطبيقات هي من نفس المصدر — `https://workspace.example.com/apps/analytics` تستدعي `https://workspace.example.com/apps/mail` — لذا لا يوجد إعداد DNS أو CORS أو - Network accessline 679
- المكالمات عبر التطبيقات هي من نفس المصدر — `https://workspace.example.com/apps/analytics` تستدعي `https://workspace.example.com/apps/mail` — لذا لا يوجد إعداد DNS أو CORS أو JWT لكل زوج. - Network accessline 687
**لا تعتمد أبدًا على `Promise` الذي يعمل بنظام "أطلق وانسى" الذي يدوم لفترة أطول من الاستجابة.** تعمل الوظائف بدون خادم (Netlify وVercel وAWS Lambda وCloud Run) على تجميد لحظة مسح نص الاستجابة - في بعض الأحيان قبل اكتمال مصافحة TCP لـ `fetc - Network accessline 725
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 726
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 727
"https://analytics.example.com", - Access to secretsline 729
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/actions-defining.mdx
- Network accessline 224
يقبل `schema` أي مكتبة متوافقة مع [Standard Schema](https://standardschema.dev):
packages/core/docs/content/locales/ar-SA/add-ui-to-your-agent.mdx
- Network accessline 27
الصق عنوان بطاقة A2A v1.0، واختر بيانات Entra. لخدمة إلى خدمة استخدم النطاق `https://ai.azure.com/.default` ودور **Foundry Agent Consumer**. - Network accessline 45
يمكنك تصفح [Global A2A Registry](https://www.a2a-registry.org). السجل للاكتشاف فقط: راجع البطاقة والمضيف ومتطلبات المصادقة ثم أضف الوكيل صراحة قبل استخدام أي سر.
packages/core/docs/content/locales/ar-SA/agent-mentions.mdx
packages/core/docs/content/locales/ar-SA/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/agent-web-surfaces.mdx
- Network accessline 169
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 170
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 171
siteUrl: "https://example.com", - Network accessline 196
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/locales/ar-SA/audit-log.mdx
- Network accessline 110
يعيد **`export-audit-ocsf`** مسار المؤسسة على هيئة أحداث [OCSF](https://schema.ocsf.io) 1.9.0 API Activity (`class_uid` 6003)، من الأقدم إلى الأحدث، ليسحبها نظام SIEM بشكل دوري. للمالكين والمسؤولين فقط، ويقرأ أحداث `org` و`admins` نفسها الت
packages/core/docs/content/locales/ar-SA/authentication.mdx
- Network accessline 8
تستخدم التطبيقات الأصلية للوكيل [Better Auth](https://better-auth.com) للمصادقة من خلال تصميم الحساب أولاً. يقوم المستخدمون بإنشاء حساب عند الزيارة الأولى ويحصلون على هوية حقيقية من اليوم الأول. - Access to secretsline 194
[Credential-related source omitted] - Access to secretsline 199
[Credential-related source omitted] - Access to secretsline 203
[Credential-related source omitted] - Access to secretsline 208
[Credential-related source omitted] - Access to secretsline 210
[Credential-related source omitted] - Access to secretsline 232
[Credential-related source omitted] - Access to secretsline 236
[Credential-related source omitted] - Network accessline 249
https://mail.agent-native.com/mcp - Access to secretsline 472
[Credential-related source omitted] - Access to secretsline 474
[Credential-related source omitted] - Access to secretsline 476
[Credential-related source omitted] - Access to secretsline 477
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/automations.mdx
packages/core/docs/content/locales/ar-SA/aws-amplify.mdx
- Access to secretsline 110
[Credential-related source omitted] - Network accessline 127
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 187
[Credential-related source omitted] - Network accessline 201
أنشئ التطبيق في [وحدة تحكم Amplify](https://console.aws.amazon.com/amplify/): - Network accessline 235
يستغرق البناء الأول بضع دقائق، ويمكنك متابعته في وحدة تحكم Amplify. عندما ينتهي، افتح عنوان التطبيق، الذي يبدو مثل `https://main.d1m7bkiki6tdw1.amplifyapp.com`، - Network accessline 236
افتح عنوان التطبيق، الذي يبدو مثل `https://main.d1m7bkiki6tdw1.amplifyapp.com`، - Access to secretsline 287
[Credential-related source omitted] - Access to secretsline 291
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/aws-lambda.mdx
- Network accessline 36
تحتاج أيضًا إلى حساب AWS وإلى تثبيت [AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html) - Access to secretsline 110
[Credential-related source omitted] - Access to secretsline 144
[Credential-related source omitted] - Access to secretsline 162
[Credential-related source omitted] - Access to secretsline 208
[Credential-related source omitted] - Access to secretsline 216
[Credential-related source omitted] - Network accessline 240
يطبع الأمر الأول قيمة `FunctionUrl`، التي تبدو مثل `https://abc123.lambda-url.us-east-1.on.aws/`. انسخها لاستخدامها في الخطوة 7. يسمح - Network accessline 241
`https://abc123.lambda-url.us-east-1.on.aws/`. انسخها لاستخدامها في الخطوة 7. يسمح - Network accessline 259
أنشئ توزيع CloudFront في [وحدة تحكم CloudFront](https://console.aws.amazon.com/cloudfront/v4/home): - Network accessline 260
[وحدة تحكم CloudFront](https://console.aws.amazon.com/cloudfront/v4/home): - Network accessline 273
في علامة التبويب **Origins** الخاصة بالتوزيع، أنشئ أصلًا. في **Origin domain**، الصق اسم المضيف من عنوان URL للدالة من الخطوة 5، بدون `https://` أو الشرطة المائلة - Network accessline 274
الصق اسم المضيف من عنوان URL للدالة من الخطوة 5، بدون `https://` أو الشرطة المائلة - Network accessline 308
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 310
[Credential-related source omitted] - Access to secretsline 312
[Credential-related source omitted] - Access to secretsline 328
[Credential-related source omitted] - Dangerous shell operationline 330
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/locales/ar-SA/azure-static-web-apps.mdx
packages/core/docs/content/locales/ar-SA/blueprint-installer.mdx
- Network accessline 68
agent-native add <kind> <name> # print a curated blueprint agent-native add <kind> <https://docs…> # البحث والتكامل من URL - Network accessline 69
agent-native add <kind> <https://docs…> # البحث والتكامل من URL - Network accessline 95
agent-native add provider https://docs.example.com/api | claude
packages/core/docs/content/locales/ar-SA/cli-adapters.mdx
- Access to secretsline 95
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/locales/ar-SA/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">Component</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 125
optional: true, description: "كل شيء من [خيارات useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) باستثناء `queryKey` و`queryFn`، التي يضبطها الخطاف نفسه.", - Network accessline 126
description: "كل شيء من [خيارات useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) باستثناء `queryKey` و`queryFn`، التي يضبطها الخطاف نفسه.", }, - Network accessline 127
"كل شيء من [خيارات useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) باستثناء `queryKey` و`queryFn`، التي يضبطها الخطاف نفسه.", - Network accessline 137
يُقصد استخدام هذا الخطاف مع actions للقراءة فقط. إنه غلاف رقيق حول دالة [useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery). - Network accessline 173
optional: true, description: "كل شيء من [خيارات useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) باستثناء `mutationFn`، التي يضبطها الخطاف نفسه.", - Network accessline 174
description: "كل شيء من [خيارات useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) باستثناء `mutationFn`، التي يضبطها الخطاف نفسه.", fields: [ - Network accessline 175
"كل شيء من [خيارات useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) باستثناء `mutationFn`، التي يضبطها الخطاف نفسه.", - Network accessline 202
يُقصد استخدام هذا الخطاف مع actions التي تغيّر البيانات، مثل الإنشاء أو التحديث أو الإرسال. إنه غلاف رقيق حول دالة [useMutation()](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation).
packages/core/docs/content/locales/ar-SA/cloneable-saas.mdx
- Network accessline 46
https://github.com/owner/repo - Network accessline 61
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 62
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 70
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 71
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 133
هل تريد إدراج قالبك؟ [أرسل قالب مجتمع](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) - Access to secretsline 287
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/cloudflare.mdx
- Access to secretsline 98
[Credential-related source omitted] - Network accessline 115
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 163
[Credential-related source omitted] - Network accessline 205
وتنظيف عمليات التشغيل المنقطعة، جميعها عبر عملية مسح تعمل مرة كل دقيقة. يضيف البناء [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) - Network accessline 206
البناء [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) بالتعبير `* * * * *` إلى ملف `wrangler.json` المُولَّد، ويحتفظ بأي Cron Triggers - Network accessline 207
[Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/)
packages/core/docs/content/locales/ar-SA/code-agents-ui.mdx
- Network accessline 204
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 205
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 357
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 358
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard"
packages/core/docs/content/locales/ar-SA/create-with-an-agent.mdx
- Network accessline 14
- وكيل ذكاء اصطناعي قادر على تشغيل أوامر الطرفية، مثل Claude Code أو Codex أو Cursor. الوكلاء الذين يقتصرون على المحادثة، مثل ChatGPT أو Claude على الويب، يمكنهم المساعدة أيضًا. راجع [الوكلاء الذين لا يستطيعون تشغيل الأوامر](#chat-only-agen - Network accessline 15
- [Node.js](https://nodejs.org) 22.22 أو أحدث. - Network accessline 37
5. شغّل التطبيق بالأمر `pnpm dev` واتركه يعمل في الخلفية. 6. اقرأ عنوان URL المحلي الذي يطبعه `pnpm dev`. تحقق من أنه يستجيب، على سبيل المثال بالأمر `curl -I <url>`. لا تفتح - Network accessline 38
6. اقرأ عنوان URL المحلي الذي يطبعه `pnpm dev`. تحقق من أنه يستجيب، على سبيل المثال بالأمر `curl -I <url>`. لا تفتح متصفحًا ولا تستخدم أدوات متصفح بلا واجهة. سأفتح التطبيق - Network accessline 39
يستجيب، على سبيل المثال بالأمر `curl -I <url>`. لا تفتح - Network accessline 57
- مجلد تطبيق جديد، باسم `my-app` ما لم تغيّره، يحتوي على `actions/` و`app/` وملف `AGENTS.md`. - تشغيل `pnpm dev` من دون أخطاء. - عنوان URL محلي، وهو عادةً `http://localhost:8080`، تحقق الوكيل من أنه يستجيب. - Network accessline 59
- عنوان URL محلي، وهو عادةً `http://localhost:8080`، تحقق الوكيل من أنه يستجيب.
packages/core/docs/content/locales/ar-SA/creating-templates.mdx
packages/core/docs/content/locales/ar-SA/decisions.mdx
- Network accessline 18
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 19
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 20
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/locales/ar-SA/deno-deploy.mdx
- Network accessline 24
تحتاج أيضًا إلى حساب Deno Deploy وإلى تثبيت [Deno](https://docs.deno.com/runtime/getting_started/installation/) - Access to secretsline 97
[Credential-related source omitted] - Network accessline 114
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 163
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/deployment-environment-variables.mdx
- Network accessline 29
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | تخطَّ DDL المخطط عند الطلب في عمليات النشر التي لا يغطيها التخطي التلقائي لـ serverless. يتطلب أن يكون مخططك مُهاجَرًا بالفعل. راجع [DDL المخطط عند الطلب](/docs/deployment#skip-ensure-tables). - Network accessline 31
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | قائمة سماح مفصولة بفواصل خاصة برد اتصال المؤسسة فقط لمصادر معاينة موثوقة دقيقة (على سبيل المثال، `https://0123456789abcdef01234567--content.netlify.app`). اضبطها فقط في عملية نشر رد الاتصال ال - Network accessline 45
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 46
| `BETTER_AUTH_SECRET` | سلسلة عشوائية لا تقل عن 32 حرفًا. توقّع ملفات تعريف ارتباط الجلسة وهي الاحتياطي لـ `OAUTH_STATE_SECRET` والقيم المشفَّرة المحلية للتطبيق. إنها أيضًا مرشح قراءة قديم لنص خزنة مشتركة مشفَّر موجو - Network accessline 47
| `BETTER_AUTH_URL` | تجاوز اختياري لعنوان URL العام لهذا التطبيق (مثل `https://mail.example.com`). يستنتجه الإطار من الطلب، والقالب المعروف، وسياق منصة الاستضافة عندما يكون غير مضبوط. - Access to secretsline 48
[Credential-related source omitted] - Access to secretsline 57
[Credential-related source omitted] - Access to secretsline 67
[Credential-related source omitted] - Access to secretsline 69
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/deployment.mdx
packages/core/docs/content/locales/ar-SA/dispatch.mdx
- Network accessline 116
يمكن أن يكون Dispatch هو موصل MCP الفردي للوكلاء الخارجيين: قم بإضافة `https://dispatch.agent-native.com/mcp` مرة واحدة في Claude أو ChatGPT أو Codex أو Cursor، وسيصل تخويل واحد إلى كل تطبيق مساحة عمل ممنوح بدلاً من موصل واحد لكل تطبيق. راج - Network accessline 125
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p
packages/core/docs/content/locales/ar-SA/docker.mdx
- Dangerous shell operationline 69
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 71
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 111
[Credential-related source omitted] - Access to secretsline 166
[Credential-related source omitted] - Network accessline 183
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/locales/ar-SA/embedding-sdk.mdx
- Network accessline 222
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 223
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 224
url="https://assets.agent-native.com/picker" - Network accessline 277
getMcpUrl("https://assets.agent-native.com"); - Network accessline 278
getA2AUrl("https://assets.agent-native.com"); - Network accessline 280
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 281
"https://assets.agent-native.com", - Network accessline 302
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 303
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 304
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 386
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 388
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 390
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 391
hostOrigin: "https://app.example.com", - Network accessline 394
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 395
hostOrigin: "https://app.example.com", - Network accessline 399
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 400
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 401
{ hostOrigin: "https://app.example.com" }, - Network accessline 405
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 406
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 407
{ hostOrigin: "https://app.example.com" }, - Network accessline 410
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 411
hostOrigin: "https://app.example.com", - Network accessline 464
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 465
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 466
hostOrigin: "https://app.example.com",
packages/core/docs/content/locales/ar-SA/extensions.mdx
- Network accessline 238
| `dbQuery(sql, args)` | القراءة من SQL (يتم تحديد نطاقها تلقائيًا للمستخدم) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` | Legacy - Network accessline 239
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` | اضغط عل - Network accessline 240
| `extensionFetch(url, options)` | اضغط على APIs الخارجية من خلال وكيل آمن به أسرار | `extensionFetch('https://api.github.com/user')` | - Network accessline 268
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 269
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/locales/ar-SA/external-agents.mdx
- Network accessline 111
https://dispatch.agent-native.com/mcp - Network accessline 122
https://mail.agent-native.com/mcp - Network accessline 126
يحتوي كل تطبيق مستضاف أيضًا على صفحة مساعدة على `https://<app>/mcp/connect` مع URL القابل للنسخ و - Network accessline 127
`https://<app>/mcp/connect` مع URL القابل للنسخ و - Network accessline 225
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 269
يقوم `skills` CLI بتثبيت ملفات `SKILL.md` فقط؛ لا يزال عملاء MCP المحليون تحتاج إلى موصل مثل `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 270
تحتاج إلى موصل مثل `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 276
تحديد العميل الافتراضي هو كافة العملاء المحليين المعتمدين؛ أضف `--client codex` أو `--client claude-code` أو هدفًا محددًا آخر لتضييق نطاق الإعداد. يعرض المضيفون المضمنون (ChatGPT ودردشة Claude على الويب وسطح المكتب والهاتف، بالإضافة إلى علا - Network accessline 282
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 297
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 300
يبحث `reconnect` عن أي إدخال تكوين MCP ينتهي URL بـ `/mcp` للمضيف المحدد والعميل المحدد (المطابقة بواسطة URL بغض النظر عن اسم الموصل)، ثم يقوم بتحديث مادة المصادقة أو استبدالها دون لمس skills المثبت أو إعادة تشغيل تدفق التثبيت الكامل. تمرير - Network accessline 308
بالنسبة لعملاء MCP الذين لا يمكنهم إضافة OAuth URL عن بعد مباشرةً، افتح التطبيق في متصفحك واستخدم تكلفة **الاتصال** (المقدمة في `https://<app>/mcp/connect`). أثناء تسجيل الدخول، انقر فوق **الاتصال / التفويض**. تمنحك الصفحة إما رابطًا - Network accessline 314
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 315
[Credential-related source omitted] - Network accessline 316
"url": "https://mail.agent-native.com/mcp", - Network accessline 332
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 333
https://dispatch.agent-native.com/mcp - Network accessline 336
هذا هو نفس الإدخال URL فقط الذي يكتبه لك `npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code`. ثم قم بتشغيل `/mcp` في رمز Claude واختر **Authenticate**. يكتشف العميل المصادقة من تحدي `401 WWW-Authen - Network accessline 339
https://dispatch.agent-native.com/mcp - Network accessline 480
npx @agent-native/core@latest Connect https://plan.agent-native.com --مخطوطة العميل --الكتالوج الكامل - Network accessline 531
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 532
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Access to secretsline 752
[Credential-related source omitted] - Access to secretsline 768
[Credential-related source omitted] - Access to secretsline 774
[Credential-related source omitted] - Network accessline 780
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 781
[Credential-related source omitted] - Network accessline 782
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 783
[Credential-related source omitted] - Network accessline 805
يعيد `connect dev` كتابة نفس أسماء خوادم MCP المستقرة (`agent-native-mail`، `agent-native-calendar`، وما إلى ذلك) إلى بوابة التطوير المحلية البطيئة، لذلك لا تتغير أسماء الأدوات. يقوم بعمل نسخة احتياطية لإدخالات الإنتاج الحالية في `~/.agent- - Network accessline 832
- قم بتوصيل وكيلك الخاص بـ Dispatch باستخدام `npx @agent-native/core@latest connect https://dispatch.agent-native.com`؛ استخدم التطبيق المباشر URL فقط عندما تريد تطبيقًا واحدًا معزولًا. - Access to secretsline 852
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/file-uploads.mdx
- Network accessline 134
سجّل موفرًا مخصصًا فقط لنوع مختلف من الواجهة الخلفية. يقوم هذا المثال بالتحميل عبر واجهة HTTP عامة باستخدام `fetch()`: - Network accessline 135
المثال بالتحميل عبر واجهة HTTP عامة باستخدام `fetch()`: - Network accessline 152
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 153
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 154
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/locales/ar-SA/getting-started.mdx
packages/core/docs/content/locales/ar-SA/harness-agents.mdx
- Network accessline 111
يمكن أن يعمل Agent-Native كعميل [ACP](https://agentclientprotocol.com) (العميل
packages/core/docs/content/locales/ar-SA/key-concepts.mdx
- Network accessline 331
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 332
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 333
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/locales/ar-SA/koyeb.mdx
packages/core/docs/content/locales/ar-SA/mcp-apps.mdx
- Network accessline 23
تضيف OpenAI Plugin Extensions نقاط تشغيل إلى جانب أدوات MCP التي يستدعيها النموذج. عند تفعيل تطبيقات MCP المضمنة، تعلن أداة `open_app` المشتركة عن نقطة دخول في الشريط الجانبي العام وأخرى في اللوحة الجانبية للمحادثة. يستدعيها ChatGPT باستخدا - Network accessline 261
تشمل المضيفات التي تعرض MCP Apps داخل المحادثة دردشة Claude على الويب وسطح المكتب والهاتف، بالإضافة إلى علامة تبويب [Code](https://github.com/anthropics/claude-code/issues/103809) في Claude Desktop؛ ChatGPT; VS Code GitHub Copilot (1.109+); - Network accessline 263
استخدم الرابط العميق البديل في طرفية Claude Code وCowork وCode على claude.ai وفي Codex CLI. [لا يعلن Cowork عن دعم MCP Apps حتى الآن، وقد يؤدي استخدامه على الموصل نفسه إلى اختفاء التطبيقات المضمنة في Chat حتى تعيد توصيل الموصل.](https://git - Access to secretsline 269
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/mcp-clients.mdx
- Network accessline 148
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 149
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 178
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 179
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 180
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 226
[Credential-related source omitted] - Network accessline 271
يستخدم إعداد Atlassian المسبق نقطة نهاية OAuth 2.1 الخاصة بـ Atlassian Rovo MCP، `https://mcp.atlassian.com/v1/mcp/authv2`، ويوفّر أدوات Jira وConfluence. - Network accessline 272
`https://mcp.atlassian.com/v1/mcp/authv2`، ويوفّر أدوات Jira وConfluence. - Network accessline 279
Rovo MCP server → Domains**، وتمكين أذونات **Read** و**Write** و**Search** المطلوبة ضمن **Permissions**. يجب أن يكون النطاق هو المضيف الذي يعمل عليه تطبيق agent-native؛ لا تنسخ `https://builder.io/**` إلى عملية نشر أخرى إلا إذا - Network accessline 280
المطلوبة ضمن **Permissions**. يجب أن يكون النطاق هو المضيف الذي يعمل عليه تطبيق agent-native؛ لا تنسخ `https://builder.io/**` إلى عملية نشر أخرى إلا إذا كان ذلك هو أصلها الفعلي. راجع [دليل إعداد Rovo MCP من Atlassian](https://developer.atla - Network accessline 281
تطبيق agent-native؛ لا تنسخ `https://builder.io/**` إلى عملية نشر أخرى إلا إذا - Network accessline 282
كان ذلك هو أصلها الفعلي. راجع [دليل إعداد Rovo MCP من Atlassian](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 283
و[عناصر تحكم المسؤول](https://support.atlassian.com/security-and-access-policies/docs/control-atlassian-rovo-mcp-server-settings/). - Network accessline 294
1. **اسم الخادم** — تسمية قصيرة للرجوع إليها (على سبيل المثال، "zapier"، "slack-tools"). 2. **URL** — نقطة النهاية HTTPS التي قدمها لك موفر خادم MCP (على سبيل المثال، `https://mcp.zapier.com/s/abc123/mcp`). يتم العثور على هذا عادةً في لوحة - Network accessline 295
2. **URL** — نقطة النهاية HTTPS التي قدمها لك موفر خادم MCP (على سبيل المثال، `https://mcp.zapier.com/s/abc123/mcp`). يتم العثور على هذا عادةً في لوحة تحكم الموفر أو مستندات التكامل. - Network accessline 310
[Credential-related source omitted] - Network accessline 402
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 403
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 404
"url": "https://example.com/mcp", - Network accessline 442
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com
packages/core/docs/content/locales/ar-SA/mcp-protocol.mdx
- Access to secretsline 31
[Credential-related source omitted] - Network accessline 108
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 117
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 118
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 119
"url": "https://mail.example.com/mcp", - Network accessline 138
POST https://your-app.example.com/mcp - Access to secretsline 143
[Credential-related source omitted] - Access to secretsline 254
[Credential-related source omitted] - Network accessline 262
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 268
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 269
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 384
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 393
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 394
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 395
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/locales/ar-SA/messaging.mdx
- Network accessline 81
1. انتقل إلى **[api.slack.com/apps](https://api.slack.com/apps)** وانقر على **إنشاء تطبيق جديد** → **من البداية**. أطلق عليه اسمًا (على سبيل المثال، "الوكيل") واختر مساحة العمل الخاصة بك. - Network accessline 97
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 120
https://your-clips.example.com/api/slack/unfurl - Access to secretsline 134
[Credential-related source omitted] - Network accessline 166
اضبط OAuth Redirect URL على `https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`، - Network accessline 167
`https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`، - Network accessline 168
وEvent Subscriptions Request URL على `https://your-app.example.com/_agent-native/integrations/slack/webhook`، وInteractivity Request URL على - Network accessline 169
`https://your-app.example.com/_agent-native/integrations/slack/webhook`، - Network accessline 170
وInteractivity Request URL على `https://your-app.example.com/_agent-native/integrations/slack/interactions`. اشترك في `app_home_opened` و`app_context_changed` و`app_mention` - Network accessline 171
`https://your-app.example.com/_agent-native/integrations/slack/interactions`. - Access to secretsline 174
[Credential-related source omitted] - Network accessline 266
[Credential-related source omitted] - Access to secretsline 266
[Credential-related source omitted] - Access to secretsline 271
[Credential-related source omitted] - Access to secretsline 368
[Credential-related source omitted] - Network accessline 382
1. افتح تيليجرام وأرسل رسالة **[@BotFather](https://t.me/BotFather)**. - Network accessline 389
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 408
1. قم بالتسجيل في **[resend.com](https://resend.com)**. الطبقة المجانية كافية للبدء. - Network accessline 415
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 422
[Credential-related source omitted] - Network accessline 429
1. قم بالتسجيل في **[sendgrid.com](https://sendgrid.com)**. - Network accessline 437
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 442
[Credential-related source omitted] - Network accessline 466
1. انتقل إلى **[Meta Developer Portal](https://developers.facebook.com/)**، وانقر على **إنشاء تطبيق**، واختر نوع **الأعمال**. - Access to secretsline 473
[Credential-related source omitted] - Network accessline 479
رد الاتصال URL: https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Access to secretsline 690
[Credential-related source omitted] - Access to secretsline 691
[Credential-related source omitted] - Access to secretsline 801
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/multi-app-workspace.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 100
[Credential-related source omitted] - Network accessline 245
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 246
"urls": { "design": "https://design.example.test" } - Network accessline 247
"design": "https://design.example.test" - Access to secretsline 276
[Credential-related source omitted] - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Network accessline 325
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 326
# https://your-agents.com/mail/* → apps/mail - Network accessline 327
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 328
# https://your-agents.com/forms/* → apps/forms
packages/core/docs/content/locales/ar-SA/native-chat-ui.mdx
- Network accessline 448
لخريطة واجهات API القديمة والجديدة وخطوات النقل، راجع [دليل نقل دردشة AgentKit](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md).
packages/core/docs/content/locales/ar-SA/netlify.mdx
packages/core/docs/content/locales/ar-SA/node-js.mdx
packages/core/docs/content/locales/ar-SA/notifications.mdx
- Network accessline 192
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 193
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 194
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 211
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 212
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 213
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/locales/ar-SA/observability.mdx
- Network accessline 507
يحتوي المجلد [`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel) في المستودع على نسخة أكمل من هذه الإضافة، مع نقاط نهاية لكل إشارة والتحقق من المُصدِّر. وهي غير منشورة على npm، لذا ابدأ من الإضافة أعلاه بدلً
packages/core/docs/content/locales/ar-SA/onboarding.mdx
- Access to secretsline 187
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/plan-plugin.mdx
- Network accessline 20
- **ثلاث مهارات** — `/visual-plan` (نقطة الدخول الأساسية)، و`/visual-recap`، و`/visualize-repo`. - **موصل Plan MCP** — مسجَّل في التطبيق المستضاف على `https://plan.agent-native.com` (نقطة نهاية MCP هي `https://plan.agent-native.com/mcp`، وا - Network accessline 21
- **موصل Plan MCP** — مسجَّل في التطبيق المستضاف على `https://plan.agent-native.com` (نقطة نهاية MCP هي `https://plan.agent-native.com/mcp`، واسم الخادم `plan`). - Network accessline 97
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 122
- **استخدم متصفحًا قائمًا على Chromium.** يحظر Safari صفحة Plan المستضافة عبر HTTPS من قراءة جسر localhost على `http://127.0.0.1` (محتوى مختلط/شبكة - Network accessline 123
HTTPS من قراءة جسر localhost على `http://127.0.0.1` (محتوى مختلط/شبكة - Network accessline 226
- `--client codex|claude-code|claude-code-cli|cowork|all` — أي الوكلاء المحليين تُكتَب تهيئة MCP لهم (الافتراضي `all`). - `--no-connect` — سجّل الموصل دون مصادقة؛ شغّل `npx @agent-native/core@latest connect https://plan.agent-native.com --c - Network accessline 227
- `--no-connect` — سجّل الموصل دون مصادقة؛ شغّل `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` لاحقًا، أو اختر `--client` أضيق. - Network accessline 282
بعد التثبيت، **ابدأ محادثة Codex جديدة** حتى تُحمَّل المهارات وأدوات MCP في الجلسة. يشحن المكوّن الإضافي موصلاً قائمًا على URL فقط (`[mcp_servers.plan]` ← `https://plan.agent-native.com/mcp`)؛ ويشغّل - Network accessline 283
الجلسة. يشحن المكوّن الإضافي موصلاً قائمًا على URL فقط (`[mcp_servers.plan]` ← `https://plan.agent-native.com/mcp`)؛ ويشغّل `codex mcp login plan` تدفق OAuth. يعمل مسار الـ CLI العام أعلاه أيضًا مع - Network accessline 284
(`[mcp_servers.plan]` ← `https://plan.agent-native.com/mcp`)؛ ويشغّل - Network accessline 292
**عمليات التثبيت الأقدم:** إذا كان التكوين لديك لا يزال يحتوي على إدخال `agent-native-plans` يشير إلى الـ URL نفسه، فإن تشغيل `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 293
`agent-native-plans` يشير إلى الـ URL نفسه، فإن تشغيل `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` لـ Codex، أو الأمر نفسه مع `--client` المستهدف لديك، يوحّده إلى الاسم - Network accessline 294
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`
packages/core/docs/content/locales/ar-SA/pr-visual-recap.mdx
- Access to secretsline 203
[Credential-related source omitted] - Access to secretsline 204
[Credential-related source omitted] - Access to secretsline 205
[Credential-related source omitted] - Network accessline 223
بالنسبة لـ DeepSeek، اضبط `VISUAL_RECAP_AGENT=openai-compatible` و`VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1` - Network accessline 224
و`VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1` - Access to secretsline 225
[Credential-related source omitted] - Access to secretsline 243
[Credential-related source omitted] - Access to secretsline 261
[Credential-related source omitted] - Network accessline 272
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 287
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 299
[Credential-related source omitted] - Access to secretsline 300
[Credential-related source omitted] - Network accessline 306
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | متغير، لمتغيّر النسخ فقط (يتجاهله المستدعي القابل لإعادة الاستخدام). اضبطه على `false` لإيقاف إنشاء الملخص عند كل دفع. عندها يُنشأ الملخص فق - Network accessline 308
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | سر. فقط عند استضافة تطبيق Plans ذاتيًا على مصدر مختلف. - Network accessline 351
لتثبيته، انسخ الملف من [BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Network accessline 352
[BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Access to secretsline 354
[Credential-related source omitted] - Access to secretsline 569
[Credential-related source omitted] - Access to secretsline 635
[Credential-related source omitted] - Access to secretsline 636
[Credential-related source omitted] - Access to secretsline 637
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/privacy-and-data-rights.mdx
- Network accessline 59
- [نص GDPR](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [المفوضية الأوروبية: التعامل مع طلبات الأفراد](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_ - Network accessline 60
- [المفوضية الأوروبية: التعامل مع طلبات الأفراد](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [النائب العام في كاليفورنيا: CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [لوائح CCPA](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/locales/ar-SA/real-time-collaboration.mdx
- Network accessline 669
يبلّغ `GET /_agent-native/health` عن النتيجة، وعملية الحل نفسها تنفّذ التسجيل، لذا فإن أمر curl واحدًا يفعّل ويؤكّد في آنٍ معًا: - Network accessline 670
فإن أمر curl واحدًا يفعّل ويؤكّد في آنٍ معًا: - Network accessline 673
curl -s https://your-app.example.com/_agent-native/health | jq .realtime
packages/core/docs/content/locales/ar-SA/recurring-jobs.mdx
packages/core/docs/content/locales/ar-SA/render.mdx
packages/core/docs/content/locales/ar-SA/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- لا تستخدم مطلقًا `innerHTML` أو `eval()` أو `document.write()` - Access to secretsline 393
[Credential-related source omitted] - Access to secretsline 409
[Credential-related source omitted] - Access to secretsline 415
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/server-database.mdx
- Network accessline 8
يخزّن كل تطبيق أصلي للوكيل حالته في PostgreSQL. تقرأ واجهة المستخدم والوكيل الجداول نفسها وتكتبان إليها عبر Actions نفسها، باستخدام عميل [Drizzle ORM](https://orm.drizzle.team) نفسه. تستخدم المزامنة الحية SSE مع الاستطلاع كخيار احتياطي، لذل - Network accessline 81
يُشغّل [PGlite](https://pglite.dev/) قاعدة بيانات Postgres مبنية بتقنية WASM داخل العملية نفسها. يتيح لك اكتشاف مشكلات المخطط الخاصة بـ Postgres فقط قبل النشر، مع إبقاء الإعداد بسيطًا مثل الإعداد الافتراضي لـ PGlite. وهو، مثل PGlite، تخزين
packages/core/docs/content/locales/ar-SA/server-middleware.mdx
- Network accessline 106
لمزيد من أنماط البرمجيات الوسيطة — المنطق الشرطي، والعمليات غير المتزامنة، واعتراض الاستجابة — راجع [وثائق البرمجيات الوسيطة في Nitro](https://nitro.build/guide/routing#middleware) و[وثائق H3](https://h3.unjs.io). - Network accessline 149
- [**المسارات**](/docs/server-routes): المسارات المخصصة المحمية بواسطة هذه البرمجية الوسيطة - [**الأمان**](/docs/security): تحديد نطاق البيانات، وحراس الوصول، والنموذج الكامل للمصادقة - [**وثائق البرمجيات الوسيطة في Nitro**](https://nitro.b - Network accessline 151
- [**وثائق البرمجيات الوسيطة في Nitro**](https://nitro.build/guide/routing#middleware): واجهة برمجة التطبيقات الكاملة للبرمجيات الوسيطة وأنماطها
packages/core/docs/content/locales/ar-SA/server-overview.mdx
- Network accessline 206
[Nitro](https://nitro.build) هو مجموعة أدوات الخادم التي تُشغّل طبقة HTTP الخاصة بالتطبيق الأصلي للوكيل. معظم ما ستُكوّنه في `server/` هو Nitro: المسارات المستندة إلى الملفات، وإضافات بدء التشغيل، والبرمجيات الوسيطة. - Network accessline 214
واجهة الطلب/الاستجابة الأساسية هي [H3](https://h3.unjs.io). عندما ترى `defineEventHandler`، أو `createError`، أو `getQuery` في مسارات مخصصة، فذلك هو H3. تُعد [وثائق Nitro](https://nitro.build/guide) و[وثائق H3](https://h3.unjs.io) المرجع لأ
packages/core/docs/content/locales/ar-SA/server-plugins.mdx
- Network accessline 139
يمكن للإضافات استخدام أي من خطافات دورة حياة Nitro: `request`، و`beforeResponse`، و`close`، وغيرها. تغطي [وثائق إضافات Nitro](https://nitro.build/guide/plugins) واجهة برمجة الخطافات الكاملة وعقد الإضافة بالتفصيل. - Network accessline 160
- [**المسارات**](/docs/server-routes): مسارات الملفات المخصصة ومتى تستخدمها - [**كتابة تعليمات الوكيل**](/docs/writing-agent-instructions): كيفية صياغة `systemPrompt` في إضافة دردشة الوكيل - [**وثائق إضافات Nitro**](https://nitro.build/guid - Network accessline 162
- [**وثائق إضافات Nitro**](https://nitro.build/guide/plugins): واجهة برمجة الإضافات الكاملة وخطافات دورة الحياة
packages/core/docs/content/locales/ar-SA/server-routes.mdx
- Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/skills-guide.mdx
- Network accessline 315
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 316
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 317
"url": "https://assets.agent-native.com", - Network accessline 318
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/locales/ar-SA/template-analytics.mdx
packages/core/docs/content/locales/ar-SA/template-assets.mdx
- Network accessline 115
عرض توضيحي مباشر: [assets.agent-native.com](https://assets.agent-native.com). - Network accessline 464
<EmbeddedApp url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 465
url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 501
تحتوي مهارة تطبيق الأصول على معرف التطبيق `assets` وتستضيف MCP URL `https://assets.agent-native.com/mcp`. - Network accessline 502
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/locales/ar-SA/template-brain.mdx
- Network accessline 155
عرض توضيحي مباشر: [brain.agent-native.com](https://brain.agent-native.com).
packages/core/docs/content/locales/ar-SA/template-calendar-developers.mdx
- Network accessline 28
يتطلب هذا تثبيت [Node.js](https://nodejs.org) الإصدار 22.22 أو أحدث و - Network accessline 29
[pnpm](https://pnpm.io) أولًا. راجع [بدء الاستخدام](/docs/getting-started) - Network accessline 39
يُفتح التطبيق على `http://localhost:8082`. إذا لم يُفتح تلقائيًا، فزُر هذا - Network accessline 49
1. **أنشئ مشروع Google Cloud**، إذا لم يكن لديك مشروع جاهز للاستخدام بالفعل: في [Google Cloud Console](https://console.cloud.google.com/)، انقر على - Network accessline 50
في [Google Cloud Console](https://console.cloud.google.com/)، انقر على - Network accessline 59
4. **أنشئ بيانات الاعتماد**: ضمن **APIs & Services → Credentials**، أنشئ معرّف عميل OAuth من نوع "Web application"، وأضف `http://localhost:8082/_agent-native/google/callback` كعنوان URI - Network accessline 60
معرّف عميل OAuth من نوع "Web application"، وأضف `http://localhost:8082/_agent-native/google/callback` كعنوان URI لإعادة توجيه معتمد (8082 هو منفذ التطوير الافتراضي للتقويم). - Network accessline 61
`http://localhost:8082/_agent-native/google/callback` كعنوان URI - Access to secretsline 77
[Credential-related source omitted] - Network accessline 80
احصل على Client ID وSecret من [تطبيق Zoom OAuth](https://marketplace.zoom.us/)
packages/core/docs/content/locales/ar-SA/template-calendar-features.mdx
packages/core/docs/content/locales/ar-SA/template-calendar-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/template-calendar.mdx
packages/core/docs/content/locales/ar-SA/template-chat-developers.mdx
- Network accessline 30
يتطلب هذا [Node.js](https://nodejs.org) الإصدار 22.22 أو أحدث. إذا لم يكن
packages/core/docs/content/locales/ar-SA/template-chat-integrations.mdx
- Access to secretsline 22
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/template-chat.mdx
packages/core/docs/content/locales/ar-SA/template-clips-developers.mdx
- Network accessline 24
يتطلب هذا تثبيت [Node.js](https://nodejs.org) بإصدار 22.22 أو أحدث - Network accessline 25
و[pnpm](https://pnpm.io) أولاً. راجع [بدء الاستخدام](/docs/getting-started) - Network accessline 321
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 322
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 323
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/template-clips-embed.mdx
- Network accessline 54
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 73
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 74
[Credential-related source omitted] - Network accessline 76
const response = await fetch( - Network accessline 191
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 199
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/locales/ar-SA/template-clips-features.mdx
- Network accessline 175
كاميرا يمكن سحبها، ونص مباشر أثناء التسجيل، كل ذلك دون الحاجة إلى ترك علامة تبويب متصفح مفتوحة. نزّله من [clips.agent-native.com/download](https://clips.agent-native.com/download)، - Network accessline 176
تبويب متصفح مفتوحة. نزّله من [clips.agent-native.com/download](https://clips.agent-native.com/download)، متوفر لـ macOS وWindows وLinux. - Network accessline 177
[clips.agent-native.com/download](https://clips.agent-native.com/download)، - Network accessline 303
[Credential-related source omitted] - Network accessline 304
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/template-clips-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/template-clips.mdx
packages/core/docs/content/locales/ar-SA/template-content.mdx
- Network accessline 16
### [جرّبه الآن →](https://content.agent-native.com)
packages/core/docs/content/locales/ar-SA/template-design-developers.mdx
packages/core/docs/content/locales/ar-SA/template-design-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted] - Network accessline 48
https://design.agent-native.com/mcp - Network accessline 52
الوصول. يضيف Claude Code العنوان نفسه، ثم يشغّل `/mcp` للمصادقة. أما Codex وعملاء CLI أو stdio الآخرون فيشغّلون بدلاً من ذلك `npx @agent-native/core@latest connect https://design.agent-native.com`، الذي - Network accessline 53
وعملاء CLI أو stdio الآخرون فيشغّلون بدلاً من ذلك `npx @agent-native/core@latest connect https://design.agent-native.com`، الذي يفتح علامة تبويب في المتصفح للتفويض ويصدر رمزًا مميزًا قابلاً للإلغاء خاصًا - Network accessline 54
@agent-native/core@latest connect https://design.agent-native.com`، الذي - Network accessline 68
فلا تستمر في إعادة المحاولة. شغّل `/mcp` في Claude Code وأعد الاتصال، أو شغّل `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` من أي طرفية. كلاهما يعيد المصادقة دون - Network accessline 69
شغّل `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` من أي طرفية. كلاهما يعيد المصادقة دون إعادة تثبيت أي شيء. - Network accessline 70
https://design.agent-native.com` من أي طرفية. كلاهما يعيد المصادقة دون
packages/core/docs/content/locales/ar-SA/template-design.mdx
packages/core/docs/content/locales/ar-SA/template-dispatch.mdx
packages/core/docs/content/locales/ar-SA/template-forms-agent.mdx
- Network accessline 40
- **"انشر هذا النموذج وأعطني الرابط."** يصبح النموذج مباشرًا وتحصل على عنوان URL العام له فورًا. - **"أنشئ نموذج تعليقات عملاء مجهولًا وأعطني الرابط العام."** يُنشر النموذج مجهولًا منذ البداية، بحيث لا تُسجَّل أي معلومة عن المرسل أبدًا. - * - Network accessline 42
- **"اسمح بإرسال هذا النموذج فقط من https://example.com."** يبدأ النموذج برفض الإرسالات من أي مكان آخر. هذا متاح فقط بسؤال الوكيل اليوم. لا يوجد مفتاح تبديل لذلك في المنشئ.
packages/core/docs/content/locales/ar-SA/template-forms-features.mdx
- Network accessline 14
[Credential-related source omitted] - Network accessline 15
[Credential-related source omitted] - Network accessline 94
يمكن أن تتطلب الإرسالات العامة تحققًا عبر [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/) قبل قبولها. فعّل هذا بضبط مفتاح الموقع (site key) ومفتاح سري (secret key) في الإعدادات. تبقى حمايتان مفعّلتين حتى بدون Turnstil - Network accessline 110
لا يوجد مفتاح تبديل لهذا في المنشئ حاليًا. اطلب من الوكيل إعداده، على سبيل المثال "اسمح بإرسال هذا النموذج فقط من https://example.com." أعطِ الوكيل الأصل الدقيق (المخطط والنطاق، مثل `https://example.com`، دون مسار لاحق) لكل موقع ينبغي السما - Network accessline 116
[Credential-related source omitted] - Network accessline 117
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/template-forms.mdx
packages/core/docs/content/locales/ar-SA/template-mail.mdx
- Network accessline 97
عرض توضيحي مباشر: [mail.agent-native.com](https://mail.agent-native.com). - Network accessline 168
1. افتح [Google Cloud Console](https://console.cloud.google.com/) وأنشئ مشروعًا. - Network accessline 169
2. قم بتمكين **Gmail API** ضمن APIs & Services → المكتبة. 3. إنشاء بيانات اعتماد OAuth 2.0 (النوع: تطبيق ويب). أضف `http://localhost:8085/_agent-native/google/callback` كإعادة توجيه معتمدة URI. 4. انسخ معرف العميل وسر العميل إلى صفحة الإعدا - Network accessline 170
3. إنشاء بيانات اعتماد OAuth 2.0 (النوع: تطبيق ويب). أضف `http://localhost:8085/_agent-native/google/callback` كإعادة توجيه معتمدة URI.
packages/core/docs/content/locales/ar-SA/template-plan.mdx
- Network accessline 137
يثبّت هذا ثلاثة أوامر شرطة مائلة — `/visual-plan` و`/visual-recap` و`/visualize-repo` — بالإضافة إلى موصل `plan` MCP (خادم أدوات عن بُعد يتواصل معه وكيلك) على `https://plan.agent-native.com/mcp`. للاطلاع على المرجع - Network accessline 138
و`/visualize-repo` — بالإضافة إلى موصل `plan` MCP (خادم أدوات عن بُعد يتواصل معه وكيلك) على `https://plan.agent-native.com/mcp`. للاطلاع على المرجع الكامل للخيارات، ومسارات تثبيت المكوّن الإضافي/السوق لـ Claude Code وCodex، - Network accessline 139
معه وكيلك) على `https://plan.agent-native.com/mcp`. للاطلاع على المرجع - Network accessline 223
إذا أرجعت أداة من أدوات Plans في أي وقت `needs auth` أو `Unauthorized` أو `Session terminated`، فلا تستمر في إعادة المحاولة. أعِد المصادقة باستخدام `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`، - Network accessline 224
`Session terminated`، فلا تستمر في إعادة المحاولة. أعِد المصادقة باستخدام `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`، أو أعد تشغيل `/mcp` ← **Authenticate** لمضيف قادر على OAuth. ابدأ محادثة - Network accessline 225
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`،
packages/core/docs/content/locales/ar-SA/template-slides-features.mdx
packages/core/docs/content/locales/ar-SA/template-slides-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/template-slides.mdx
packages/core/docs/content/locales/ar-SA/toolkit-capability-packages.mdx
packages/core/docs/content/locales/ar-SA/toolkit-settings.mdx
- Access to secretsline 121
[Credential-related source omitted] - Network accessline 191
يتطلب تدفق الخدمة الذاتية من OpenAI استدعاء OAuth عبر HTTP على `127.0.0.1`، وهو عنوان loopback على جهاز المستخدم. لذلك لا تعمل التطبيقات المستضافة عن بُعد، بما فيها المواقع على `*.agent-native.com`، كما أن `localhost` ليس بديلًا مقبولًا عن - Network accessline 201
استخدم Responses API العامة وحدها للاستدلال. لا تشمل هذه الموافقة نقاط النهاية الخاصة بالخلفية في ChatGPT أو بيانات محادثاته. تصف OpenAI تدفق مشاركة الرموز بأنه إصدار معاينة، لذا قد يتغير توفر النماذج والحدود وميزات API المدعومة. راجع [نظرة
packages/core/docs/content/locales/ar-SA/tracking.mdx
- Network accessline 83
| المزود | إنف فار | | ----------------------- | --------------------------------------------------------------------------------- - Network accessline 84
| ----------------------- | ---------------------------------------------------------------------------------------------------- | | ما بعد الهوج | `POSTHOG_API_KEY` (مطلوب)، `POSTHOG_HOST` (اختياري، الإعداد الافتراضي هو `https:/ - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Network accessline 205
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 206
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 207
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 225
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 226
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 283
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 284
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 285
fetch("https://analytics.example.com/events", { - Network accessline 343
- **HTTP المجمعة** — يقوم الموفرون المضمنون بوضع الأحداث في قائمة الانتظار وتدفقها كل 10 ثوانٍ أو عندما يتراكم 50 حدثًا، أيهما يأتي أولاً. يؤدي هذا إلى تقليل الطلبات الصادرة دون فقدان البيانات. - **لا توجد تبعيات SDK** — يستخدم جميع موفري ا - Network accessline 344
- **لا توجد تبعيات SDK** — يستخدم جميع موفري الخدمة المضمنين `fetch()` الخام. لا يوجد PostHog SDK، ولا Mixpanel SDK، ولا Amplitude SDK. يحافظ على إطار العمل خفيف الوزن.
packages/core/docs/content/locales/ar-SA/turn-into-app.mdx
- Network accessline 12
- وكيل ذكاء اصطناعي يستطيع تشغيل أوامر الطرفية، مثل Claude Code وCodex وCursor. يمكن لـ Claude أو ChatGPT على الويب تسليم المهمة إلى Builder بدلاً من ذلك. راجع [أين تعمل](#hosts). - [Node.js](https://nodejs.org). يتضمن مثبّت nodejs.org الأد - Network accessline 13
- [Node.js](https://nodejs.org). يتضمن مثبّت nodejs.org الأداة `npx` التي تشغّل أوامر التثبيت أدناه. - Network accessline 18
- **على جهازك.** Node.js بإصدار 22.22 أو أحدث (تحقق بالأمر `node --version`) و[pnpm](https://pnpm.io) ضمن `PATH` لديك. إذا فشل الأمر `pnpm --version`، فشغّل `corepack enable` الذي يأتي مع Node.js. - Network accessline 59
تتطلب الاستضافة لدى Builder، من أي نوع من المضيفين، موصل Agent-Native Dispatch MCP (`https://dispatch.agent-native.com/mcp`، OAuth) وتستدعي إجراءه `start-workspace-app-creation`. لا يعمل شيء على جانبك، لذا يبقى التقرير معلّقاً أو غير متحقق - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/locales/ar-SA/updating-ui-in-production.mdx
- Network accessline 18
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment) هو سير عمل مُدار منفصل قادر على التعامل مع الكود حيث يمكن لوكيل تعديل واجهة تطبيقك في الإنتاج. استخدم Builder.io مع م
packages/core/docs/content/locales/ar-SA/upgrading-core-ui.mdx
- Access to secretsline 68
[Credential-related source omitted] - Access to secretsline 71
[Credential-related source omitted] - Network accessline 178
عنوان المصدر الأساسي: [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx). - Network accessline 179
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx).
packages/core/docs/content/locales/ar-SA/vercel.mdx
- Access to secretsline 94
[Credential-related source omitted] - Network accessline 111
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 137
[Credential-related source omitted] - Network accessline 195
تتم الأتمتة المجدولة، والأتمتة المُشغَّلة بالأحداث المنتظرة في قائمة الانتظار، وتنظيف عمليات التشغيل المنقطعة، جميعها عبر عملية مسح تعمل مرة كل دقيقة. يضيف البناء [مهمة cron في Vercel](https://vercel.com/docs/cron-jobs) تستدعي عملية - Network accessline 196
وتنظيف عمليات التشغيل المنقطعة، جميعها عبر عملية مسح تعمل مرة كل دقيقة. يضيف البناء [مهمة cron في Vercel](https://vercel.com/docs/cron-jobs) تستدعي عملية المسح على `/_agent-native/jobs/_process-sweep`. ويضيف بناء مساحة العمل مهمة - Network accessline 197
البناء [مهمة cron في Vercel](https://vercel.com/docs/cron-jobs) تستدعي عملية
packages/core/docs/content/locales/ar-SA/voice-input.mdx
packages/core/docs/content/locales/ar-SA/what-is-agent-native.mdx
packages/core/docs/content/locales/ar-SA/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Network accessline 267
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 268
[Credential-related source omitted] - Network accessline 269
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 270
[Credential-related source omitted] - Access to secretsline 271
[Credential-related source omitted] - Access to secretsline 272
[Credential-related source omitted] - Network accessline 284
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 285
[Credential-related source omitted] - Network accessline 286
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 287
[Credential-related source omitted] - Access to secretsline 288
[Credential-related source omitted] - Access to secretsline 289
[Credential-related source omitted] - Network accessline 340
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 342
APP_URL=https://workspace.example.com
packages/core/docs/content/locales/ar-SA/workspace-management.mdx
- Network accessline 90
ليس كل من يحتاج إلى إجراء تغييرات يشعر بالارتياح تجاه git. يدعم [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management) نموذج التفرع المرئي الذي يعين فروع git المو - Network accessline 140
نصائح أساسية: استخدم فرق GitHub (`@org/team`)، وليس الأفراد. يجب أن تتطلب تغييرات الإطار والإرسال دائمًا مراجعة النظام الأساسي. راجع [GitHub CODEOWNERS docs](https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-fea - Network accessline 160
ثم أضف الإجراء [actions/labeler](https://github.com/actions/labeler) - راجع README الخاص بالريبو للتعرف على سير العمل الكامل YAML. يتم تطبيق التصنيفات تلقائيًا عند فتح العلاقات العامة أو تحديثها. - Access to secretsline 233
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/a2a-protocol.mdx
- Access to secretsline 95
[Credential-related source omitted] - Network accessline 119
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 120
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 121
"url": "https://analytics.example.com", - Network accessline 316
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 352
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 353
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 354
"https://analytics.example.com", - Access to secretsline 356
[Credential-related source omitted] - Network accessline 376
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", - Network accessline 377
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", "cardUrl": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a/agentCard/v1.0", - Network accessline 378
"url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", - Network accessline 379
"cardUrl": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a/agentCard/v1.0", - Network accessline 419
[Azure AI Foundry hosted agents](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 426
Foundry kann über [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint) - Network accessline 438
Für die Client-Anmeldeinformationen eines Dienstprinzipals fordern Sie ein Entra-Token mit dem Bereich `https://ai.azure.com/.default` an und weisen Sie - Network accessline 439
Entra-Token mit dem Bereich `https://ai.azure.com/.default` an und weisen Sie - Network accessline 446
"auth": { "type": "oauth-client-credentials", "tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", - Network accessline 447
"type": "oauth-client-credentials", "tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", "clientId": "<application-id>", - Network accessline 448
"tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", - Network accessline 449
"clientId": "<application-id>", "clientSecretRef": "FOUNDRY_CLIENT_SECRET", "scope": "https://ai.azure.com/.default" - Network accessline 450
[Credential-related source omitted] - Access to secretsline 450
[Credential-related source omitted] - Network accessline 451
"scope": "https://ai.azure.com/.default" - Network accessline 458
[Gemini Enterprise kann externe A2A-Agenten registrieren](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent), - Network accessline 472
`discoveryengine.assist`. Die Registrierung eigener A2A-Agenten ist Pre-GA. Prüfen Sie daher die Verfügbarkeit im Zielprojekt, bevor Sie sie im Arbeitsbereichs-Picker anbieten. Siehe [Gemini-Enterprise-Agent aufrufen](https://docs.cloud.goo - Network accessline 473
Prüfen Sie daher die Verfügbarkeit im Zielprojekt, bevor Sie sie im Arbeitsbereichs-Picker anbieten. Siehe [Gemini-Enterprise-Agent aufrufen](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) für Ressourcen-URL und IAM- - Network accessline 474
Arbeitsbereichs-Picker anbieten. Siehe [Gemini-Enterprise-Agent aufrufen](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) - Network accessline 492
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 493
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 494
"url": "https://api.anthropic.com", - Access to secretsline 499
[Credential-related source omitted] - Access to secretsline 518
[Credential-related source omitted] - Network accessline 564
# Eine bereitgestellte App direkt per URL aufrufen und das Bearer-Token aus der Umgebung lesen agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 565
[Credential-related source omitted] - Access to secretsline 565
[Credential-related source omitted] - Access to secretsline 566
[Credential-related source omitted] - Access to secretsline 640
[Credential-related source omitted] - Access to secretsline 644
[Credential-related source omitted] - Network accessline 702
- Ein gemeinsamer `A2A_SECRET` wird zur Erstellungszeit in die Umgebung jeder App eingebunden. - App-übergreifende Aufrufe haben denselben Ursprung – `https://workspace.example.com/apps/analytics` ruft `https://workspace.example.com/apps/ma - Network accessline 703
- App-übergreifende Aufrufe haben denselben Ursprung – `https://workspace.example.com/apps/analytics` ruft `https://workspace.example.com/apps/mail` auf – es gibt also kein DNS-, CORS- oder paarweises JWT-Setup. - Network accessline 711
**Verlassen Sie sich nie darauf, dass ein Fire-and-Forget-`Promise` die Antwort überlebt.** Serverlose Funktionen (Netlify, Vercel, AWS Lambda, Cloud Run) frieren ein, sobald der Antworttext geleert wird – manchmal bevor der TCP-Handshake e - Network accessline 749
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 750
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 751
"https://analytics.example.com", - Access to secretsline 753
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/actions-defining.mdx
- Network accessline 226
`schema` akzeptiert jede mit [Standard Schema](https://standardschema.dev) kompatible Bibliothek:
packages/core/docs/content/locales/de-DE/add-ui-to-your-agent.mdx
- Network accessline 27
Fügen Sie die A2A-v1.0-Agent-Card-URL ein und wählen Sie die Entra-Zugangsdaten. Für Dienst-zu-Dienst-Zugriff verwenden Sie `https://ai.azure.com/.default` und die Rolle **Foundry Agent Consumer**. - Network accessline 45
Über die [Global A2A Registry](https://www.a2a-registry.org) können Sie entdecken. Prüfen Sie Card, Host und Authentifizierung und fügen Sie einen Agenten ausdrücklich hinzu, bevor ein Secret verwendet wird.
packages/core/docs/content/locales/de-DE/agent-mentions.mdx
packages/core/docs/content/locales/de-DE/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/agent-web-surfaces.mdx
- Network accessline 171
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 172
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 173
siteUrl: "https://example.com", - Network accessline 198
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/locales/de-DE/audit-log.mdx
- Network accessline 110
**`export-audit-ocsf`** liefert die Spur der Organisation als [OCSF](https://schema.ocsf.io)-1.9.0-API-Activity-Ereignisse (`class_uid` 6003), älteste zuerst, damit ein SIEM sie regelmäßig abrufen kann. Nur für Eigentümer und Administratore
packages/core/docs/content/locales/de-DE/authentication.mdx
- Network accessline 8
Agent-Native Apps verwenden [Better Auth](https://better-auth.com) für die Authentifizierung mit einem Account-First-Design. Benutzer erstellen beim ersten Besuch ein Konto und erhalten vom ersten Tag an eine echte Identität. - Access to secretsline 197
[Credential-related source omitted] - Access to secretsline 202
[Credential-related source omitted] - Access to secretsline 206
[Credential-related source omitted] - Access to secretsline 211
[Credential-related source omitted] - Access to secretsline 213
[Credential-related source omitted] - Access to secretsline 235
[Credential-related source omitted] - Access to secretsline 239
[Credential-related source omitted] - Network accessline 252
https://mail.agent-native.com/mcp - Access to secretsline 345
[Credential-related source omitted] - Access to secretsline 477
[Credential-related source omitted] - Access to secretsline 479
[Credential-related source omitted] - Access to secretsline 481
[Credential-related source omitted] - Access to secretsline 482
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/automations.mdx
packages/core/docs/content/locales/de-DE/aws-amplify.mdx
- Access to secretsline 121
[Credential-related source omitted] - Network accessline 140
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 203
[Credential-related source omitted] - Network accessline 217
Erstellen Sie die App in der [Amplify-Konsole](https://console.aws.amazon.com/amplify/): - Network accessline 253
Der erste Build dauert einige Minuten, und Sie können ihn in der Amplify-Konsole verfolgen. Wenn er fertig ist, öffnen Sie die Adresse der App, die etwa wie `https://main.d1m7bkiki6tdw1.amplifyapp.com` aussieht, und erstellen Sie das - Network accessline 254
verfolgen. Wenn er fertig ist, öffnen Sie die Adresse der App, die etwa wie `https://main.d1m7bkiki6tdw1.amplifyapp.com` aussieht, und erstellen Sie das erste Konto. - Network accessline 255
`https://main.d1m7bkiki6tdw1.amplifyapp.com` aussieht, und erstellen Sie das - Access to secretsline 314
[Credential-related source omitted] - Access to secretsline 318
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/aws-lambda.mdx
- Network accessline 41
Außerdem brauchen Sie ein AWS-Konto und die [AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html), - Network accessline 42
[AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html), - Access to secretsline 123
[Credential-related source omitted] - Access to secretsline 160
[Credential-related source omitted] - Access to secretsline 178
[Credential-related source omitted] - Access to secretsline 228
[Credential-related source omitted] - Access to secretsline 236
[Credential-related source omitted] - Network accessline 262
Der erste Befehl gibt die `FunctionUrl` aus, die etwa so aussieht: `https://abc123.lambda-url.us-east-1.on.aws/`. Kopieren Sie sie für Schritt 7. - Network accessline 263
`https://abc123.lambda-url.us-east-1.on.aws/`. Kopieren Sie sie für Schritt 7. - Network accessline 283
Erstellen Sie die CloudFront-Distribution in der [CloudFront-Konsole](https://console.aws.amazon.com/cloudfront/v4/home): - Network accessline 284
[CloudFront-Konsole](https://console.aws.amazon.com/cloudfront/v4/home): - Network accessline 300
Erstellen Sie auf dem Tab **Origins** der Distribution einen Origin. Fügen Sie bei **Origin domain** den Hostnamen der Funktions-URL aus Schritt 5 ein, ohne `https://` und ohne den abschließenden Schrägstrich. Setzen Sie **Protocol** auf - Network accessline 301
bei **Origin domain** den Hostnamen der Funktions-URL aus Schritt 5 ein, ohne `https://` und ohne den abschließenden Schrägstrich. Setzen Sie **Protocol** auf **HTTPS only** und **Response timeout** auf 60 Sekunden. Fügen Sie diesem Origin - Network accessline 302
`https://` und ohne den abschließenden Schrägstrich. Setzen Sie **Protocol** auf - Network accessline 340
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 342
[Credential-related source omitted] - Access to secretsline 344
[Credential-related source omitted] - Access to secretsline 362
[Credential-related source omitted] - Dangerous shell operationline 364
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/locales/de-DE/azure-static-web-apps.mdx
packages/core/docs/content/locales/de-DE/blueprint-installer.mdx
- Network accessline 68
agent-native add <kind> <name> # print a curated blueprint agent-native add <kind> <https://docs…> # Forschung und Integration von einem URL aus - Network accessline 69
agent-native add <kind> <https://docs…> # Forschung und Integration von einem URL aus - Network accessline 95
agent-native add provider https://docs.example.com/api | claude
packages/core/docs/content/locales/de-DE/cli-adapters.mdx
- Access to secretsline 99
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/locales/de-DE/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">Komponente</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 126
optional: true, description: "Alles aus den [useQuery-Optionen](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) außer `queryKey` und `queryFn`, die der Hook selbst setzt.", - Network accessline 127
description: "Alles aus den [useQuery-Optionen](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) außer `queryKey` und `queryFn`, die der Hook selbst setzt.", }, - Network accessline 128
"Alles aus den [useQuery-Optionen](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) außer `queryKey` und `queryFn`, die der Hook selbst setzt.", - Network accessline 138
Dieser Hook ist für schreibgeschützte Actions gedacht. Es ist ein dünner Wrapper um die Funktion [useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery). - Network accessline 174
optional: true, description: "Alles aus den [useMutation-Optionen](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) außer `mutationFn`, die der Hook selbst setzt.", - Network accessline 175
description: "Alles aus den [useMutation-Optionen](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) außer `mutationFn`, die der Hook selbst setzt.", fields: [ - Network accessline 176
"Alles aus den [useMutation-Optionen](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) außer `mutationFn`, die der Hook selbst setzt.", - Network accessline 203
Dieser Hook ist für Actions gedacht, die Daten ändern, wie Erstellen, Aktualisieren oder Senden. Es ist ein dünner Wrapper um die Funktion [useMutation()](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation).
packages/core/docs/content/locales/de-DE/cloneable-saas.mdx
- Network accessline 49
https://github.com/owner/repo - Network accessline 67
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 68
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 77
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 78
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 157
Möchten Sie aufgenommen werden? [Reichen Sie ein Community-Template ein](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) - Access to secretsline 312
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/cloudflare.mdx
- Access to secretsline 105
[Credential-related source omitted] - Network accessline 124
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 176
[Credential-related source omitted] - Network accessline 221
Aufräumen unterbrochener Läufe erledigt ein Sweep, der einmal pro Minute läuft. Der Build fügt der generierten `wrangler.json` einen [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) - Network accessline 222
Der Build fügt der generierten `wrangler.json` einen [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) mit `* * * * *` hinzu und behält alle Cron Triggers bei, die Sie bereits - Network accessline 223
[Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/)
packages/core/docs/content/locales/de-DE/code-agents-ui.mdx
- Network accessline 207
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 208
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 360
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 361
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard"
packages/core/docs/content/locales/de-DE/create-with-an-agent.mdx
- Network accessline 14
- Ein KI-Agent, der Terminalbefehle ausführen kann, zum Beispiel Claude Code, Codex oder Cursor. Agenten, die nur chatten, wie ChatGPT oder Claude im Web, können trotzdem helfen. Siehe [Agenten, die keine Befehle ausführen können](#chat-onl - Network accessline 15
- [Node.js](https://nodejs.org) 22.22 oder höher. - Network accessline 39
5. Starte die App mit `pnpm dev` und lass sie im Hintergrund laufen. 6. Lies die lokale URL, die `pnpm dev` ausgibt. Prüfe, ob sie antwortet, zum Beispiel mit `curl -I <url>`. Öffne keinen Browser - Network accessline 40
6. Lies die lokale URL, die `pnpm dev` ausgibt. Prüfe, ob sie antwortet, zum Beispiel mit `curl -I <url>`. Öffne keinen Browser und nutze keine Headless-Browser-Tools. Ich öffne die App selbst. - Network accessline 41
antwortet, zum Beispiel mit `curl -I <url>`. Öffne keinen Browser - Network accessline 60
- Ein neuer App-Ordner, der `my-app` heißt, sofern Sie den Namen nicht geändert haben, und `actions/`, `app/` und eine Datei `AGENTS.md` enthält. - `pnpm dev` läuft ohne Fehler. - Eine lokale URL, normalerweise `http://localhost:8080`, bei - Network accessline 62
- Eine lokale URL, normalerweise `http://localhost:8080`, bei der der Agent geprüft hat, dass sie antwortet.
packages/core/docs/content/locales/de-DE/creating-templates.mdx
- Access to secretsline 487
[Credential-related source omitted] - Network accessline 532
[Community-Vorlage einreichen](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml), wenn das Repository bereit ist. Community-Einträge bleiben vom kleineren First-Party-Katalog getrennt und werden von ihren
packages/core/docs/content/locales/de-DE/cross-app-sso.mdx
- Access to secretsline 65
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/decisions.mdx
- Network accessline 18
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 19
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 20
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/locales/de-DE/deno-deploy.mdx
- Network accessline 26
Außerdem brauchen Sie ein Deno-Deploy-Konto und eine Installation von [Deno](https://docs.deno.com/runtime/getting_started/installation/) auf Ihrem - Network accessline 27
[Deno](https://docs.deno.com/runtime/getting_started/installation/) auf Ihrem - Access to secretsline 106
[Credential-related source omitted] - Network accessline 125
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 179
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/deployment-environment-variables.mdx
- Network accessline 31
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | Überspringt On-Demand-Schema-DDL auf Bereitstellungen, die das automatische serverlose Überspringen nicht abdeckt. Setzt voraus, dass Ihr Schema bereits migriert ist. Siehe [On-Demand-Schema-D - Network accessline 33
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | Nur für den Unternehmens-Callback bestimmte, kommagetrennte Zulassungsliste exakter vertrauenswürdiger Vorschau-Ursprünge (zum Beispiel `https://0123456789abcdef01234567--content.netlify.app`) - Network accessline 48
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 49
| `BETTER_AUTH_SECRET` | Zufällige Zeichenfolge mit 32+ Zeichen. Signiert Sitzungscookies und ist der Fallback für `OAUTH_STATE_SECRET` und app-lokale verschlüsselte Werte. Sie ist außerdem ein Legacy-Lesekandidat für - Network accessline 50
| `BETTER_AUTH_URL` | Optionale Überschreibung des öffentlichen Ursprungs für diese App (z. B. `https://mail.example.com`). Das Framework leitet ihn aus Anfrage-, bekanntem-Template- und Hosting-Plattform-Kontext a - Access to secretsline 51
[Credential-related source omitted] - Access to secretsline 61
[Credential-related source omitted] - Access to secretsline 86
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/deployment.mdx
packages/core/docs/content/locales/de-DE/dispatch.mdx
- Network accessline 116
Dispatch kann der einzelne MCP-Connector für externe Agenten sein: Fügen Sie `https://dispatch.agent-native.com/mcp` einmal in Claude, ChatGPT, Codex oder Cursor hinzu, und eine Autorisierung erreicht jede gewährte Workspace-App anstelle ei - Network accessline 125
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p
packages/core/docs/content/locales/de-DE/docker.mdx
- Dangerous shell operationline 75
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 77
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 119
[Credential-related source omitted] - Access to secretsline 179
[Credential-related source omitted] - Network accessline 199
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/locales/de-DE/embedding-sdk.mdx
- Network accessline 224
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 225
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 226
url="https://assets.agent-native.com/picker" - Network accessline 279
getMcpUrl("https://assets.agent-native.com"); - Network accessline 280
getA2AUrl("https://assets.agent-native.com"); - Network accessline 282
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 283
"https://assets.agent-native.com", - Network accessline 304
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 305
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 306
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 388
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 390
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 392
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 393
hostOrigin: "https://app.example.com", - Network accessline 396
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 397
hostOrigin: "https://app.example.com", - Network accessline 401
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 402
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 403
{ hostOrigin: "https://app.example.com" }, - Network accessline 407
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 408
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 409
{ hostOrigin: "https://app.example.com" }, - Network accessline 412
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 413
hostOrigin: "https://app.example.com", - Network accessline 466
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 467
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 468
hostOrigin: "https://app.example.com",
packages/core/docs/content/locales/de-DE/extensions.mdx
- Network accessline 241
| `dbQuery(sql, args)` | Lesen aus SQL (automatisch auf den Benutzer beschränkt) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` | Lega - Network accessline 242
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` | Tref - Network accessline 243
| `extensionFetch(url, options)` | Treffen Sie externe APIs über einen sicheren Proxy mit Geheimnissen | `extensionFetch('https://api.github.com/user')` | - Network accessline 271
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 272
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/locales/de-DE/external-agents.mdx
- Network accessline 112
https://dispatch.agent-native.com/mcp - Network accessline 123
https://mail.agent-native.com/mcp - Network accessline 127
Jede gehostete App verfügt auch über eine Hilfsseite unter `https://<app>/mcp/connect` mit dem kopierbaren URL und - Network accessline 128
`https://<app>/mcp/connect` mit dem kopierbaren URL und - Network accessline 235
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 279
Der rohe `skills` CLI installiert nur `SKILL.md`-Dateien; lokale MCP-Clients noch benötigen einen Anschluss wie `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 280
benötigen einen Anschluss wie `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 286
Die Standard-Clientauswahl umfasst alle unterstützten lokalen Clients. Fügen Sie `--client codex`, `--client claude-code` oder ein anderes spezifisches Ziel hinzu, um das Setup einzugrenzen. Inline-Hosts (ChatGPT, Claude-Chat im Web, auf de - Network accessline 292
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 307
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 310
`reconnect` findet jeden MCP-Konfigurationseintrag, dessen URL auf `/mcp` für den angegebenen Host und ausgewählten Client endet (Übereinstimmung mit URL unabhängig vom Connector-Namen), und aktualisiert oder ersetzt dann das Authentifizier - Network accessline 318
Für MCP-Clients, die keinen Remote-OAuth URL direkt hinzufügen können, öffnen Sie die App in Ihrem Browser und nutzen Sie das **Connect**-Angebot (bereitgestellt unter `https://<app>/mcp/connect`). Klicken Sie im angemeldeten Zustand - Network accessline 324
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 325
[Credential-related source omitted] - Network accessline 326
"url": "https://mail.agent-native.com/mcp", - Network accessline 342
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 343
https://dispatch.agent-native.com/mcp - Network accessline 346
Dies ist derselbe Nur-URL-Eintrag, den `npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code` für Sie schreibt. Führen Sie dann `/mcp` im Claude-Code aus und wählen Sie **Authentifizieren**. Der Clien - Network accessline 349
https://dispatch.agent-native.com/mcp - Network accessline 491
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex --full-catalog - Network accessline 542
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 543
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Access to secretsline 766
[Credential-related source omitted] - Access to secretsline 782
[Credential-related source omitted] - Access to secretsline 788
[Credential-related source omitted] - Network accessline 794
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 795
[Credential-related source omitted] - Network accessline 796
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 797
[Credential-related source omitted] - Network accessline 819
`connect dev` schreibt die gleichen stabilen MCP-Servernamen (`agent-native-mail`, `agent-native-calendar` usw.) in das lokale Dev-Lazy-Gateway um, sodass sich die Toolnamen nicht ändern. Es sichert die aktuellen Produktionseinträge in `~/. - Access to secretsline 833
[Credential-related source omitted] - Network accessline 846
- Verbinden Sie Ihren eigenen Agenten mit Dispatch mit `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; Verwenden Sie eine direkte App URL nur, wenn Sie eine isolierte App wünschen. - Access to secretsline 866
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/file-uploads.mdx
- Network accessline 134
Registrieren Sie einen benutzerdefinierten Anbieter nur für eine andere Art von Backend. Dieses Beispiel lädt über eine generische HTTP-API mit `fetch()` hoch: - Network accessline 136
`fetch()` hoch: - Network accessline 153
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 154
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 155
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/locales/de-DE/getting-started.mdx
- Network accessline 23
- [Node.js](https://nodejs.org) 22.22 oder höher - Network accessline 24
- [pnpm](https://pnpm.io), installiert und über Ihren `PATH` verfügbar - Network accessline 156
Starten Sie den Entwicklungsserver neu, damit er die neue Route erkennt, und öffnen Sie anschließend `http://localhost:<port>/hello`. - Network accessline 157
öffnen Sie anschließend `http://localhost:<port>/hello`.
packages/core/docs/content/locales/de-DE/harness-agents.mdx
- Network accessline 111
Agent-Native kann als [ACP](https://agentclientprotocol.com) (Agent Client) fungieren
packages/core/docs/content/locales/de-DE/key-concepts.mdx
- Network accessline 332
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 333
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 334
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/locales/de-DE/koyeb.mdx
packages/core/docs/content/locales/de-DE/mcp-apps.mdx
- Network accessline 23
OpenAI Plugin Extensions ergänzen die von Modellen aufgerufenen MCP-Tools um Startpunkte. Wenn Inline-MCP-Apps aktiviert sind, kündigt das gemeinsame Tool `open_app` einen globalen Einstiegspunkt in der Seitenleiste und einen im Seitenberei - Network accessline 261
MCP Apps werden in folgenden Hosts inline angezeigt: Claude-Chat im Web, auf dem Desktop und auf Mobilgeräten sowie die Registerkarte [Code](https://github.com/anthropics/claude-code/issues/103809) in Claude Desktop; ChatGPT; VS Code GitHub - Network accessline 263
Verwenden Sie im Claude Code-Terminal, in Cowork, in Code auf claude.ai und in der Codex CLI den Deep-Link-Fallback. [Cowork deklariert noch keine Unterstützung für MCP Apps; bei Nutzung desselben Connectors können Inline-Apps in Chat versc - Access to secretsline 269
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/mcp-clients.mdx
- Network accessline 148
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 149
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 178
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 179
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 180
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 226
[Credential-related source omitted] - Network accessline 274
Das Atlassian-Preset verwendet den OAuth-2.1-Endpunkt von Atlassian Rovo MCP, `https://mcp.atlassian.com/v1/mcp/authv2`, und stellt Tools für Jira und - Network accessline 275
`https://mcp.atlassian.com/v1/mcp/authv2`, und stellt Tools für Jira und - Network accessline 285
Berechtigungen **Read**, **Write** und **Search** aktivieren. Die Domain muss der Host sein, auf dem die agent-native-App ausgeführt wird. Übernehmen Sie `https://builder.io/**` nicht für eine andere Bereitstellung, sofern dies nicht - Network accessline 286
Host sein, auf dem die agent-native-App ausgeführt wird. Übernehmen Sie `https://builder.io/**` nicht für eine andere Bereitstellung, sofern dies nicht deren tatsächlicher Ursprung ist. Weitere Informationen finden Sie im - Network accessline 287
`https://builder.io/**` nicht für eine andere Bereitstellung, sofern dies nicht - Network accessline 288
deren tatsächlicher Ursprung ist. Weitere Informationen finden Sie im [Rovo-MCP-Einrichtungsleitfaden von Atlassian](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) und in den [Administratorsteuerelementen](https://s - Network accessline 289
[Rovo-MCP-Einrichtungsleitfaden von Atlassian](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 290
und in den [Administratorsteuerelementen](https://support.atlassian.com/security-and-access-policies/docs/control-atlassian-rovo-mcp-server-settings/). - Network accessline 303
1. **Servername** – eine kurze Bezeichnung für Ihre eigene Referenz (z. B. „zapier“, „slack-tools“). 2. **URL** – der HTTPS-Endpunkt, den Ihnen der MCP-Serveranbieter gegeben hat (z. B. `https://mcp.zapier.com/s/abc123/mcp`). Dies finden Si - Network accessline 304
2. **URL** – der HTTPS-Endpunkt, den Ihnen der MCP-Serveranbieter gegeben hat (z. B. `https://mcp.zapier.com/s/abc123/mcp`). Dies finden Sie normalerweise im Dashboard oder in den Integrationsdokumenten des Anbieters. - Network accessline 319
[Credential-related source omitted] - Network accessline 411
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 412
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 413
"url": "https://example.com/mcp", - Network accessline 451
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com
packages/core/docs/content/locales/de-DE/mcp-protocol.mdx
- Access to secretsline 31
[Credential-related source omitted] - Network accessline 108
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 117
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 118
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 119
"url": "https://mail.example.com/mcp", - Network accessline 138
POST https://your-app.example.com/mcp - Access to secretsline 143
[Credential-related source omitted] - Access to secretsline 256
[Credential-related source omitted] - Network accessline 264
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 270
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 271
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 386
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 395
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 396
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 397
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/locales/de-DE/messaging.mdx
- Network accessline 81
1. Gehen Sie zu **[api.slack.com/apps](https://api.slack.com/apps)** und klicken Sie auf **Neue App erstellen** → **Von Grund auf neu**. Benennen Sie es (z. B. „Agent“) und wählen Sie Ihren Arbeitsbereich aus. - Network accessline 97
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 120
https://your-clips.example.com/api/slack/unfurl - Access to secretsline 134
[Credential-related source omitted] - Network accessline 170
Setzen Sie die OAuth Redirect URL auf `https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 171
`https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 172
die Event Subscriptions Request URL auf `https://your-app.example.com/_agent-native/integrations/slack/webhook` und die Interactivity Request URL auf - Network accessline 173
`https://your-app.example.com/_agent-native/integrations/slack/webhook` und die - Network accessline 174
Interactivity Request URL auf `https://your-app.example.com/_agent-native/integrations/slack/interactions`. Abonnieren Sie `app_home_opened`, `app_context_changed`, `app_mention`, - Network accessline 175
`https://your-app.example.com/_agent-native/integrations/slack/interactions`. - Access to secretsline 178
[Credential-related source omitted] - Network accessline 274
[Credential-related source omitted] - Access to secretsline 274
[Credential-related source omitted] - Access to secretsline 279
[Credential-related source omitted] - Access to secretsline 376
[Credential-related source omitted] - Network accessline 390
1. Telegramm öffnen und **[@BotFather](https://t.me/BotFather)** senden. - Network accessline 397
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 416
1. Melden Sie sich unter **[resend.com](https://resend.com)** an. Die kostenlose Stufe reicht für den Einstieg aus. - Network accessline 423
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 430
[Credential-related source omitted] - Network accessline 437
1. Melden Sie sich unter **[sendgrid.com](https://sendgrid.com)** an. - Network accessline 445
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 450
[Credential-related source omitted] - Network accessline 474
1. Gehen Sie zu **[Meta Developer Portal](https://developers.facebook.com/)**, klicken Sie auf **App erstellen** und wählen Sie den Typ **Unternehmen** aus. - Access to secretsline 481
[Credential-related source omitted] - Network accessline 487
Rückruf URL: https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Access to secretsline 698
[Credential-related source omitted] - Access to secretsline 699
[Credential-related source omitted] - Access to secretsline 809
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/multi-app-workspace.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 100
[Credential-related source omitted] - Network accessline 245
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 246
"urls": { "design": "https://design.example.test" } - Network accessline 247
"design": "https://design.example.test" - Access to secretsline 276
[Credential-related source omitted] - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Network accessline 325
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 326
# https://your-agents.com/mail/* → apps/mail - Network accessline 327
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 328
# https://your-agents.com/forms/* → apps/forms
packages/core/docs/content/locales/de-DE/native-chat-ui.mdx
- Network accessline 453
Eine Übersicht der alten und neuen APIs mit Migrationsschritten finden Sie im [AgentKit-Chat-Migrationsleitfaden](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md).
packages/core/docs/content/locales/de-DE/netlify.mdx
packages/core/docs/content/locales/de-DE/node-js.mdx
packages/core/docs/content/locales/de-DE/notifications.mdx
- Network accessline 192
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 193
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 194
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 211
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 212
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 213
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/locales/de-DE/observability.mdx
- Network accessline 507
[`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel) im Repository enthält eine umfassendere Version dieses Plugins mit Endpunkten pro Signal und Exporter-Validierung. Das Paket ist nicht auf npm veröffentlic
packages/core/docs/content/locales/de-DE/onboarding.mdx
- Access to secretsline 187
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/plan-plugin.mdx
- Network accessline 22
- **Drei Skills** — `/visual-plan` (der kanonische Einstiegspunkt), `/visual-recap` und `/visualize-repo`. - **Den Plan-MCP-Connector** — registriert gegen die gehostete App unter `https://plan.agent-native.com` (MCP-Endpunkt `https://plan. - Network accessline 23
- **Den Plan-MCP-Connector** — registriert gegen die gehostete App unter `https://plan.agent-native.com` (MCP-Endpunkt `https://plan.agent-native.com/mcp`, Servername `plan`). - Network accessline 101
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 131
- **Verwenden Sie einen Chromium-Browser.** Safari blockiert, dass die gehostete HTTPS-Plan-Seite die lokale `http://127.0.0.1`-Bridge liest - Network accessline 132
gehostete HTTPS-Plan-Seite die lokale `http://127.0.0.1`-Bridge liest - Network accessline 251
- `--no-connect` — registriert den Connector ohne Authentifizierung; führen Sie später `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 252
führen Sie später `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` aus, oder wählen Sie einen engeren `--client`. - Network accessline 253
`npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 324
die Skills und MCP-Tools in die Sitzung geladen werden. Das Plugin liefert einen reinen URL-Connector (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`); `codex mcp login plan` führt den - Network accessline 325
liefert einen reinen URL-Connector (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`); `codex mcp login plan` führt den OAuth-Flow aus. Der obige universelle CLI-Weg funktioniert auch für - Network accessline 326
`https://plan.agent-native.com/mcp`); `codex mcp login plan` führt den - Network accessline 337
`agent-native-plans` hat, der auf dieselbe URL zeigt, konsolidiert das Ausführen von `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 338
Ausführen von `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` für Codex, oder derselbe Befehl mit Ihrem Ziel-`--client`, diesen zum - Network accessline 339
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`
packages/core/docs/content/locales/de-DE/pr-visual-recap.mdx
- Access to secretsline 240
[Credential-related source omitted] - Access to secretsline 241
[Credential-related source omitted] - Access to secretsline 242
[Credential-related source omitted] - Network accessline 278
Für DeepSeek setzen Sie `VISUAL_RECAP_AGENT=openai-compatible`, `VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`, - Network accessline 279
`VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`, - Access to secretsline 281
[Credential-related source omitted] - Access to secretsline 319
[Credential-related source omitted] - Access to secretsline 350
[Credential-related source omitted] - Network accessline 364
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 382
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 395
[Credential-related source omitted] - Access to secretsline 396
[Credential-related source omitted] - Network accessline 402
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | Variable, nur für die Copy-Variante (der wiederverwendbare Aufrufer ignoriert sie). Auf `false` setzen, um nicht mehr bei jedem Push e - Network accessline 404
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | Secret. Nur wenn Sie die Plans-App unter einem anderen Ursprung selbst hosten. - Network accessline 453
Um sie zu installieren, kopieren Sie die Datei von [BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Network accessline 454
[BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Access to secretsline 457
[Credential-related source omitted] - Access to secretsline 705
[Credential-related source omitted] - Access to secretsline 776
[Credential-related source omitted] - Access to secretsline 777
[Credential-related source omitted] - Access to secretsline 778
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/privacy-and-data-rights.mdx
- Network accessline 59
- [DSGVO-Text](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [Europäische Kommission: Anfragen von Einzelpersonen](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-in - Network accessline 60
- [Europäische Kommission: Anfragen von Einzelpersonen](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [California Attorney General: CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [CCPA-Regelungen](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/locales/de-DE/real-time-collaboration.mdx
- Network accessline 691
curl -s https://your-app.example.com/_agent-native/health | jq .realtime
packages/core/docs/content/locales/de-DE/recurring-jobs.mdx
packages/core/docs/content/locales/de-DE/render.mdx
packages/core/docs/content/locales/de-DE/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- Verwenden Sie niemals `innerHTML`, `eval()` oder `document.write()` - Access to secretsline 395
[Credential-related source omitted] - Access to secretsline 411
[Credential-related source omitted] - Access to secretsline 417
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/server-database.mdx
- Network accessline 8
Jede agentennative App speichert ihren Zustand in PostgreSQL. UI und Agent lesen und schreiben dieselben Tabellen über dieselben Actions, mit demselben [Drizzle ORM](https://orm.drizzle.team)-Client. Die Live-Synchronisierung verwendet SSE - Network accessline 81
[PGlite](https://pglite.dev/) betreibt eine In-Process-WASM-Postgres-Datenbank. Damit können Sie Postgres-spezifische Schemaprobleme erkennen, bevor Sie deployen, während die Einrichtung so einfach bleibt wie beim PGlite-Standard. Wie PGlit
packages/core/docs/content/locales/de-DE/server-middleware.mdx
- Network accessline 104
Weitere Middleware-Muster — bedingte Logik, asynchrone Operationen, Response-Interception — finden Sie in der [Nitro-Middleware-Dokumentation](https://nitro.build/guide/routing#middleware) und der [H3-Dokumentation](https://h3.unjs.io). - Network accessline 147
- [**Routen**](/docs/server-routes): benutzerdefinierte Routen, die durch diese Middleware geschützt werden - [**Security**](/docs/security): data scoping, access guards und das vollständige Auth-Modell - [**Nitro-Middleware-Dokumentation** - Network accessline 149
- [**Nitro-Middleware-Dokumentation**](https://nitro.build/guide/routing#middleware): vollständige Middleware-API und -Muster
packages/core/docs/content/locales/de-DE/server-overview.mdx
- Network accessline 204
[Nitro](https://nitro.build) ist das Server-Toolkit, das die agentennative HTTP-Schicht antreibt. Das meiste, was Sie in `server/` konfigurieren, ist Nitro: dateibasierte Routen, Startup-Plugins und Middleware. - Network accessline 212
Die zugrunde liegende Request/Response-API ist [H3](https://h3.unjs.io). Wenn Sie in benutzerdefinierten Routen `defineEventHandler`, `createError` oder `getQuery` sehen, ist das H3. Die [Nitro-Dokumentation](https://nitro.build/guide) und
packages/core/docs/content/locales/de-DE/server-plugins.mdx
- Network accessline 140
Plugins können jeden der Lifecycle-Hooks von Nitro verwenden: `request`, `beforeResponse`, `close` und weitere. Die [Nitro-Plugin-Dokumentation](https://nitro.build/guide/plugins) behandelt die vollständige Hook-API und den Plugin-Vertrag i - Network accessline 162
- [**Routen**](/docs/server-routes): benutzerdefinierte Dateirouten und wann sie sinnvoll sind - [**Writing Agent Instructions**](/docs/writing-agent-instructions): wie man den `systemPrompt` im Agent-Chat-Plugin gestaltet - [**Nitro-Plugin - Network accessline 164
- [**Nitro-Plugin-Dokumentation**](https://nitro.build/guide/plugins): die vollständige Plugin-API und Lifecycle-Hooks
packages/core/docs/content/locales/de-DE/server-routes.mdx
- Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/skills-guide.mdx
- Network accessline 315
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 316
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 317
"url": "https://assets.agent-native.com", - Network accessline 318
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/locales/de-DE/template-analytics.mdx
packages/core/docs/content/locales/de-DE/template-assets.mdx
- Network accessline 115
Live-Demo: [assets.agent-native.com](https://assets.agent-native.com). - Network accessline 464
<EmbeddedApp url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 465
url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 501
Der Assets-App-Skill hat die App-ID `assets` und wird als MCP URL gehostet `https://assets.agent-native.com/mcp`. - Network accessline 502
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/locales/de-DE/template-brain.mdx
- Network accessline 159
Live-Demo: [brain.agent-native.com](https://brain.agent-native.com).
packages/core/docs/content/locales/de-DE/template-calendar-developers.mdx
- Network accessline 24
Dafür benötigen Sie zunächst [Node.js](https://nodejs.org) 22.22 oder neuer und [pnpm](https://pnpm.io). Wenn Sie noch keines von beiden installiert haben oder keine LLM-Verbindung (Builder.io, ein Anthropic/OpenAI-Schlüssel oder lokales Ol - Network accessline 32
Die App öffnet sich unter `http://localhost:8082`. Wenn sie sich nicht automatisch öffnet, rufen Sie diese URL in Ihrem Browser auf. Kalender funktioniert vollständig auch ohne verbundenes Google-Konto. Sie stoßen nur auf einen Fehler, wenn - Network accessline 38
1. **Ein Google Cloud-Projekt erstellen**, falls Sie noch keines zur Verwendung haben: Klicken Sie in der [Google Cloud Console](https://console.cloud.google.com/) oben auf der Seite auf das Projekt-Dropdown, dann auf **Neues Projekt**. Geb - Network accessline 39
2. **Die API aktivieren**: Gehen Sie bei ausgewähltem Projekt zu **APIs & Dienste → Bibliothek**, suchen Sie nach „Google Calendar API“, öffnen Sie es und klicken Sie auf **Aktivieren**. 3. **Den Zustimmungsbildschirm konfigurieren**: Richt - Network accessline 40
3. **Den Zustimmungsbildschirm konfigurieren**: Richten Sie unter **APIs & Dienste → OAuth-Zustimmungsbildschirm** einen ein („Extern“ reicht für die Entwicklung). 4. **Anmeldedaten erstellen**: Erstellen Sie unter **APIs & Dienste → Anmeld - Network accessline 41
4. **Anmeldedaten erstellen**: Erstellen Sie unter **APIs & Dienste → Anmeldedaten** eine OAuth-Client-ID vom Typ „Webanwendung“ und fügen Sie `http://localhost:8082/_agent-native/google/callback` als autorisierte Weiterleitungs-URI hinzu ( - Access to secretsline 52
[Credential-related source omitted] - Network accessline 55
Holen Sie sich eine Client-ID und ein Secret von einer [Zoom OAuth-App](https://marketplace.zoom.us/) (Zoom Marketplace → Develop → Build App → OAuth), mit `/_agent-native/zoom/callback` als Weiterleitungs-URI, und starten Sie dann `pnpm de
packages/core/docs/content/locales/de-DE/template-calendar-features.mdx
packages/core/docs/content/locales/de-DE/template-calendar-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/template-calendar.mdx
packages/core/docs/content/locales/de-DE/template-chat-developers.mdx
- Network accessline 25
Dafür benötigen Sie [Node.js](https://nodejs.org) 22.22 oder neuer. Wenn `pnpm` nicht in Ihrem `PATH` liegt, führen Sie `corepack enable` in der generierten App aus. Wenn Sie Node.js noch nicht installiert haben, lesen Sie [Erste Schritte](
packages/core/docs/content/locales/de-DE/template-chat-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/template-chat.mdx
packages/core/docs/content/locales/de-DE/template-clips-developers.mdx
- Network accessline 22
Dafür müssen zuerst [Node.js](https://nodejs.org) 22.22 oder neuer und [pnpm](https://pnpm.io) installiert sein. Siehe [Erste Schritte](/docs/getting-started), wenn Sie noch keines von beiden installiert haben oder keine LLM-Verbindung (Bui - Network accessline 309
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 310
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 311
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/template-clips-embed.mdx
- Network accessline 40
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 53
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 54
[Credential-related source omitted] - Network accessline 56
const response = await fetch( - Network accessline 145
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 152
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/locales/de-DE/template-clips-features.mdx
- Network accessline 117
Die Desktop-App ist für die reibungsloseste alltägliche Aufnahme gebaut: Bildschirm- und Webcam-Aufnahme mit einer ziehbaren Kamerablase, Live-Transkription während der Aufnahme, alles ohne einen offenen Browser-Tab. Laden Sie sie unter [cl - Network accessline 212
[Credential-related source omitted] - Network accessline 213
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/template-clips-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/template-clips.mdx
packages/core/docs/content/locales/de-DE/template-content.mdx
- Network accessline 12
### [Jetzt ausprobieren →](https://content.agent-native.com)
packages/core/docs/content/locales/de-DE/template-design-developers.mdx
packages/core/docs/content/locales/de-DE/template-design-integrations.mdx
- Network accessline 55
https://design.agent-native.com/mcp - Network accessline 60
dieselbe URL hinzu und führt dann `/mcp` aus, um sich zu authentifizieren. Codex und andere CLI- oder stdio-Clients führen stattdessen `npx @agent-native/core@latest connect https://design.agent-native.com` aus, was - Network accessline 61
Codex und andere CLI- oder stdio-Clients führen stattdessen `npx @agent-native/core@latest connect https://design.agent-native.com` aus, was einen Browser-Tab zur Autorisierung öffnet und ein pro Benutzer - Network accessline 62
@agent-native/core@latest connect https://design.agent-native.com` aus, was - Network accessline 79
nicht immer wieder erneut. Führen Sie `/mcp` in Claude Code aus und verbinden Sie sich erneut, oder führen Sie `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` in einem beliebigen Terminal - Network accessline 80
verbinden Sie sich erneut, oder führen Sie `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` in einem beliebigen Terminal aus. Beide authentifizieren sich neu, ohne irgendetwas neu zu installieren. - Network accessline 81
reconnect https://design.agent-native.com` in einem beliebigen Terminal
packages/core/docs/content/locales/de-DE/template-design.mdx
packages/core/docs/content/locales/de-DE/template-dispatch.mdx
packages/core/docs/content/locales/de-DE/template-forms-agent.mdx
- Network accessline 40
- **„Veröffentliche das und gib mir den Link.“** Das Formular geht sofort live, und Sie erhalten seine öffentliche URL. - **„Erstelle ein anonymes Kundenfeedback-Formular und gib mir den öffentlichen Link.“** Das Formular wird von Anfang an - Network accessline 42
- **„Erlaube die Einreichung dieses Formulars nur von https://example.com aus.“** Das Formular beginnt, Einreichungen von überall sonst abzulehnen. Das ist heute nur über den Agenten möglich. Dafür gibt es keinen Schalter im Builder.
packages/core/docs/content/locales/de-DE/template-forms-features.mdx
- Network accessline 14
[Credential-related source omitted] - Network accessline 15
[Credential-related source omitted] - Network accessline 94
Öffentliche Einreichungen können eine Verifizierung über [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/) verlangen, bevor sie durchgelassen werden. Aktivieren Sie das, indem Sie in den Einstellungen einen Site Key und - Network accessline 110
Dafür gibt es heute keinen Schalter im Builder. Bitten Sie den Agenten, das einzurichten, zum Beispiel „erlaube die Einreichung dieses Formulars nur von https://example.com aus.“ Geben Sie dem Agenten für jede zulässige Website den genauen - Network accessline 116
[Credential-related source omitted] - Network accessline 117
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/template-forms.mdx
packages/core/docs/content/locales/de-DE/template-mail.mdx
- Network accessline 97
Live-Demo: [mail.agent-native.com](https://mail.agent-native.com). - Network accessline 168
1. Öffnen Sie [Google Cloud Console](https://console.cloud.google.com/) und erstellen Sie ein Projekt. - Network accessline 169
2. Aktivieren Sie **Gmail API** unter APIs & Services → Bibliothek. 3. Erstellen Sie OAuth 2.0-Anmeldeinformationen (Typ: Webanwendung). Fügen Sie `http://localhost:8085/_agent-native/google/callback` als autorisierte Weiterleitung URI hinz - Network accessline 170
3. Erstellen Sie OAuth 2.0-Anmeldeinformationen (Typ: Webanwendung). Fügen Sie `http://localhost:8085/_agent-native/google/callback` als autorisierte Weiterleitung URI hinzu.
packages/core/docs/content/locales/de-DE/template-plan.mdx
- Network accessline 146
und `/visualize-repo` — sowie den `plan`-MCP-Connector (ein entfernter Tool-Server, mit dem Ihr Agent kommuniziert) unter `https://plan.agent-native.com/mcp`. Die vollständige Flag-Referenz, die - Network accessline 147
Tool-Server, mit dem Ihr Agent kommuniziert) unter `https://plan.agent-native.com/mcp`. Die vollständige Flag-Referenz, die Plugin-/Marketplace-Installationswege für Claude Code und Codex sowie - Network accessline 148
`https://plan.agent-native.com/mcp`. Die vollständige Flag-Referenz, die - Network accessline 252
terminated` zurückgibt, versuchen Sie es nicht immer wieder erneut. Authentifizieren Sie sich erneut mit `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, - Network accessline 253
Authentifizieren Sie sich erneut mit `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, oder führen Sie `/mcp` → **Authentifizieren** für einen OAuth-fähigen - Network accessline 254
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`,
packages/core/docs/content/locales/de-DE/template-slides-features.mdx
packages/core/docs/content/locales/de-DE/template-slides-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/template-slides.mdx
packages/core/docs/content/locales/de-DE/toolkit-capability-packages.mdx
packages/core/docs/content/locales/de-DE/toolkit-settings.mdx
- Access to secretsline 121
[Credential-related source omitted] - Network accessline 191
Der Self-Service-Ablauf von OpenAI erfordert einen HTTP-OAuth-Callback unter `127.0.0.1`, einer Loopback-Adresse auf dem Rechner des Nutzers. Deshalb werden gehostete Apps nicht unterstützt, einschließlich der Websites unter `*.agent-native - Network accessline 201
Verwende für Inferenz ausschließlich die öffentliche Responses API. Private Backend-Endpunkte von ChatGPT und ChatGPT-Unterhaltungsdaten sind nicht Teil dieser Berechtigung. OpenAI kennzeichnet den Token-Sharing-Ablauf als Vorschau; Modellv
packages/core/docs/content/locales/de-DE/tracking.mdx
- Network accessline 83
| Anbieter | Umgebungsvariablen | | --------- | -------------------------------------------------------------------------------------------------- - Network accessline 84
| --------- | --------------------------------------------------------------------------------------------------------------- | | PostHog | `POSTHOG_API_KEY` (erforderlich), `POSTHOG_HOST` (optional, Standardeinstellung ist `https://us.i. - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Network accessline 205
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 206
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 207
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 225
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 226
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 283
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 284
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 285
fetch("https://analytics.example.com/events", { - Network accessline 343
- **Batched HTTP** – integrierte Anbieter stellen Ereignisse alle 10 Sekunden in die Warteschlange und leeren sie, oder wenn sich 50 Ereignisse ansammeln, je nachdem, was zuerst eintritt. Dadurch werden ausgehende Anfragen minimiert, ohne d - Network accessline 344
- **Keine SDK-Abhängigkeiten** – alle integrierten Anbieter verwenden rohes `fetch()`. Kein PostHog SDK, kein Mixpanel SDK, kein Amplitude SDK. Hält das Framework leichtgewichtig.
packages/core/docs/content/locales/de-DE/turn-into-app.mdx
- Network accessline 12
- Ein KI-Agent, der Terminalbefehle ausführen kann, etwa Claude Code, Codex oder Cursor. Claude oder ChatGPT im Web kann die Aufgabe stattdessen an Builder übergeben. Siehe [Wo er läuft](#hosts). - [Node.js](https://nodejs.org). Das Install - Network accessline 13
- [Node.js](https://nodejs.org). Das Installationsprogramm von nodejs.org enthält `npx`, mit dem die folgenden Installationsbefehle laufen. - Network accessline 18
- **Auf Ihrem Rechner.** Node.js 22.22 oder höher (prüfen mit `node --version`) und [pnpm](https://pnpm.io) in Ihrem `PATH`. Schlägt `pnpm --version` fehl, führen Sie `corepack enable` aus, das mit Node.js ausgeliefert wird. - Network accessline 59
Builder-Hosting benötigt von beiden Host-Arten aus den Agent-Native-Dispatch-MCP-Connector (`https://dispatch.agent-native.com/mcp`, OAuth) und ruft dessen Aktion `start-workspace-app-creation` auf. Auf Ihrer Seite läuft nichts, daher bleib - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/locales/de-DE/updating-ui-in-production.mdx
- Network accessline 25
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment)
packages/core/docs/content/locales/de-DE/upgrading-core-ui.mdx
- Access to secretsline 76
[Credential-related source omitted] - Access to secretsline 79
[Credential-related source omitted] - Network accessline 193
Kanonische Quell-URL: [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx). - Network accessline 194
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx).
packages/core/docs/content/locales/de-DE/vercel.mdx
- Access to secretsline 102
[Credential-related source omitted] - Network accessline 121
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 149
[Credential-related source omitted] - Network accessline 215
Geplante Automatisierungen, wartende Ereignis-Automatisierungen und das Aufräumen unterbrochener Läufe erledigt ein Sweep, der einmal pro Minute läuft. Der Build fügt einen [Vercel-Cron-Job](https://vercel.com/docs/cron-jobs) hinzu, - Network accessline 216
Aufräumen unterbrochener Läufe erledigt ein Sweep, der einmal pro Minute läuft. Der Build fügt einen [Vercel-Cron-Job](https://vercel.com/docs/cron-jobs) hinzu, der den Sweep unter `/_agent-native/jobs/_process-sweep` aufruft. Ein - Network accessline 217
Der Build fügt einen [Vercel-Cron-Job](https://vercel.com/docs/cron-jobs) hinzu,
packages/core/docs/content/locales/de-DE/voice-input.mdx
packages/core/docs/content/locales/de-DE/what-is-agent-native.mdx
packages/core/docs/content/locales/de-DE/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Network accessline 269
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 270
[Credential-related source omitted] - Network accessline 271
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 272
[Credential-related source omitted] - Access to secretsline 273
[Credential-related source omitted] - Access to secretsline 274
[Credential-related source omitted] - Network accessline 286
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 287
[Credential-related source omitted] - Network accessline 288
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 289
[Credential-related source omitted] - Access to secretsline 290
[Credential-related source omitted] - Access to secretsline 291
[Credential-related source omitted] - Network accessline 343
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 345
APP_URL=https://workspace.example.com
packages/core/docs/content/locales/de-DE/workspace-management.mdx
- Network accessline 90
Nicht jeder, der Änderungen vornehmen muss, ist mit Git vertraut. [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management) unterstützt ein visuelles Verzweigungsmod - Network accessline 140
Wichtige Tipps: Verwenden Sie GitHub-Teams (`@org/team`), keine Einzelpersonen. Framework- und Dispatch-Änderungen sollten immer eine Plattformüberprüfung erfordern. Informationen zur Glob-Syntax und zu Mustern mit mehreren Eigentümern find - Network accessline 160
Dann fügen Sie die Aktion [actions/labeler](https://github.com/actions/labeler) hinzu – siehe README des Repos für den vollständigen Workflow YAML. Beschriftungen werden automatisch angewendet, wenn PRs geöffnet oder aktualisiert werden. - Access to secretsline 233
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/a2a-protocol.mdx
- Access to secretsline 95
[Credential-related source omitted] - Network accessline 119
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 120
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 121
"url": "https://analytics.example.com", - Access to secretsline 164
[Credential-related source omitted] - Network accessline 316
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 352
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 353
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 354
"https://analytics.example.com", - Access to secretsline 356
[Credential-related source omitted] - Network accessline 376
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", - Network accessline 377
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", "cardUrl": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a/agentCard/v1.0", - Network accessline 378
"url": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a", - Network accessline 379
"cardUrl": "https://example.services.ai.azure.com/agents/foundry-research/endpoint/protocols/a2a/agentCard/v1.0", - Network accessline 417
[Los agentes alojados de Azure AI Foundry](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 424
Foundry puede exponer un punto de conexión A2A mediante [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint). - Network accessline 435
Para las credenciales de cliente de una entidad de servicio, solicita un token de Entra con el ámbito `https://ai.azure.com/.default` y asigna al llamador el - Network accessline 436
de Entra con el ámbito `https://ai.azure.com/.default` y asigna al llamador el - Network accessline 442
"auth": { "type": "oauth-client-credentials", "tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", - Network accessline 443
"type": "oauth-client-credentials", "tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", "clientId": "<application-id>", - Network accessline 444
"tokenUrl": "https://login.microsoftonline.com/<tenant>/oauth2/v2.0/token", - Network accessline 445
"clientId": "<application-id>", "clientSecretRef": "FOUNDRY_CLIENT_SECRET", "scope": "https://ai.azure.com/.default" - Network accessline 446
[Credential-related source omitted] - Access to secretsline 446
[Credential-related source omitted] - Network accessline 447
"scope": "https://ai.azure.com/.default" - Network accessline 454
[Gemini Enterprise puede registrar agentes A2A externos](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent), - Network accessline 468
`discoveryengine.assist`. El registro personalizado de agentes A2A es Pre-GA, así que confirme que está disponible en el proyecto de destino antes de mostrarlo en el selector del espacio de trabajo. Consulte [invocar un agente de Gemini Ent - Network accessline 469
así que confirme que está disponible en el proyecto de destino antes de mostrarlo en el selector del espacio de trabajo. Consulte [invocar un agente de Gemini Enterprise](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a - Network accessline 470
mostrarlo en el selector del espacio de trabajo. Consulte [invocar un agente de Gemini Enterprise](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) - Network accessline 488
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 489
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 490
"url": "https://api.anthropic.com", - Access to secretsline 495
[Credential-related source omitted] - Access to secretsline 514
[Credential-related source omitted] - Network accessline 560
# Llama directamente a una aplicación desplegada por URL, leyendo el token de portador desde una variable de entorno agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 561
[Credential-related source omitted] - Access to secretsline 561
[Credential-related source omitted] - Access to secretsline 562
[Credential-related source omitted] - Access to secretsline 636
[Credential-related source omitted] - Access to secretsline 640
[Credential-related source omitted] - Access to secretsline 653
[Credential-related source omitted] - Network accessline 698
- Se monta un `A2A_SECRET` compartido en el entorno de cada aplicación en el momento de la compilación. - Las llamadas entre aplicaciones tienen el mismo origen (`https://workspace.example.com/apps/analytics` llama a `https://workspace.exam - Network accessline 699
- Las llamadas entre aplicaciones tienen el mismo origen (`https://workspace.example.com/apps/analytics` llama a `https://workspace.example.com/apps/mail`), por lo que no hay configuración de DNS, CORS o JWT por par. - Network accessline 707
**Nunca confíe en que un `Promise` de tipo "disparar y olvidar" sobreviva a la respuesta.** Las funciones sin servidor (Netlify, Vercel, AWS Lambda, Cloud Run) se congelan en el momento en que se vacía el cuerpo de la respuesta, a veces inc - Network accessline 745
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 746
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 747
"https://analytics.example.com", - Access to secretsline 749
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/actions-defining.mdx
- Network accessline 226
`schema` acepta cualquier librería compatible con [Standard Schema](https://standardschema.dev):
packages/core/docs/content/locales/es-ES/add-ui-to-your-agent.mdx
- Network accessline 27
Pega la URL de la tarjeta A2A v1.0 y elige las credenciales de Entra. Para servicio a servicio usa `https://ai.azure.com/.default` y el rol **Foundry Agent Consumer**. - Network accessline 45
Explora la [Global A2A Registry](https://www.a2a-registry.org). Verifica la tarjeta, el host y la autenticación antes de añadir un agente y usar credenciales.
packages/core/docs/content/locales/es-ES/agent-mentions.mdx
packages/core/docs/content/locales/es-ES/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/agent-web-surfaces.mdx
- Network accessline 171
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 172
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 173
siteUrl: "https://example.com", - Network accessline 198
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/locales/es-ES/audit-log.mdx
- Network accessline 110
**`export-audit-ocsf`** devuelve el rastro de la organización como eventos OCSF 1.9.0 API Activity ([OCSF](https://schema.ocsf.io), `class_uid` 6003), del más antiguo al más reciente, para que un SIEM los extraiga de forma programada. Solo
packages/core/docs/content/locales/es-ES/authentication.mdx
- Network accessline 8
Las aplicaciones nativas del agente utilizan [Better Auth](https://better-auth.com) para la autenticación con un diseño que prioriza la cuenta. Los usuarios crean una cuenta en la primera visita y obtienen una identidad real desde el primer - Access to secretsline 196
[Credential-related source omitted] - Access to secretsline 201
[Credential-related source omitted] - Access to secretsline 205
[Credential-related source omitted] - Access to secretsline 210
[Credential-related source omitted] - Access to secretsline 212
[Credential-related source omitted] - Access to secretsline 234
[Credential-related source omitted] - Access to secretsline 238
[Credential-related source omitted] - Network accessline 251
https://mail.agent-native.com/mcp - Access to secretsline 476
[Credential-related source omitted] - Access to secretsline 478
[Credential-related source omitted] - Access to secretsline 480
[Credential-related source omitted] - Access to secretsline 481
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/automations.mdx
packages/core/docs/content/locales/es-ES/aws-amplify.mdx
- Access to secretsline 113
[Credential-related source omitted] - Network accessline 130
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 192
[Credential-related source omitted] - Network accessline 206
Crea la aplicación en la [consola de Amplify](https://console.aws.amazon.com/amplify/): - Network accessline 241
La primera compilación tarda unos minutos, y puedes seguirla en la consola de Amplify. Cuando termine, abre la dirección de la aplicación, que tiene un aspecto como `https://main.d1m7bkiki6tdw1.amplifyapp.com`, y crea la primera cuenta. - Network accessline 243
`https://main.d1m7bkiki6tdw1.amplifyapp.com`, y crea la primera cuenta. - Access to secretsline 295
[Credential-related source omitted] - Access to secretsline 299
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/aws-lambda.mdx
- Network accessline 38
También necesitas una cuenta de AWS y la [AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html) - Access to secretsline 114
[Credential-related source omitted] - Access to secretsline 149
[Credential-related source omitted] - Access to secretsline 167
[Credential-related source omitted] - Access to secretsline 215
[Credential-related source omitted] - Access to secretsline 223
[Credential-related source omitted] - Network accessline 248
El primer comando muestra la `FunctionUrl`, que tiene un aspecto como `https://abc123.lambda-url.us-east-1.on.aws/`. Cópiala para el paso 7. Los dos - Network accessline 249
`https://abc123.lambda-url.us-east-1.on.aws/`. Cópiala para el paso 7. Los dos - Network accessline 268
Crea la distribución de CloudFront en la [consola de CloudFront](https://console.aws.amazon.com/cloudfront/v4/home): - Network accessline 269
[consola de CloudFront](https://console.aws.amazon.com/cloudfront/v4/home): - Network accessline 283
En la pestaña **Origins** de la distribución, crea un origen. En **Origin domain**, pega el nombre de host de la URL de la función del paso 5, sin `https://` ni la - Network accessline 284
pega el nombre de host de la URL de la función del paso 5, sin `https://` ni la - Network accessline 320
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 322
[Credential-related source omitted] - Access to secretsline 324
[Credential-related source omitted] - Access to secretsline 341
[Credential-related source omitted] - Dangerous shell operationline 343
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/locales/es-ES/azure-static-web-apps.mdx
packages/core/docs/content/locales/es-ES/blueprint-installer.mdx
- Network accessline 68
agent-native add <kind> <name> # print a curated blueprint agent-native add <kind> <https://docs…> # investigar e integrar desde un URL - Network accessline 69
agent-native add <kind> <https://docs…> # investigar e integrar desde un URL - Network accessline 95
agent-native add provider https://docs.example.com/api | claude
packages/core/docs/content/locales/es-ES/cli-adapters.mdx
- Access to secretsline 99
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/locales/es-ES/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">Componente</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 126
optional: true, description: "Todo lo de [las opciones de useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) salvo `queryKey` y `queryFn`, que el hook establece por sí mismo.", - Network accessline 127
description: "Todo lo de [las opciones de useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) salvo `queryKey` y `queryFn`, que el hook establece por sí mismo.", }, - Network accessline 128
"Todo lo de [las opciones de useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) salvo `queryKey` y `queryFn`, que el hook establece por sí mismo.", - Network accessline 138
Este hook está pensado para usarse con acciones de solo lectura. Es un envoltorio ligero alrededor de la función [useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery). - Network accessline 174
optional: true, description: "Todo lo de [las opciones de useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) salvo `mutationFn`, que el hook establece por sí mismo.", - Network accessline 175
description: "Todo lo de [las opciones de useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) salvo `mutationFn`, que el hook establece por sí mismo.", fields: [ - Network accessline 176
"Todo lo de [las opciones de useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) salvo `mutationFn`, que el hook establece por sí mismo.", - Network accessline 203
Este hook está pensado para usarse con acciones que cambian datos, como crear, actualizar o enviar algo. Es un envoltorio ligero alrededor de la función [useMutation()](https://tanstack.com/query/latest/docs/framework/react/reference/useMut
packages/core/docs/content/locales/es-ES/cloneable-saas.mdx
- Network accessline 49
https://github.com/owner/repo - Network accessline 67
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 68
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 77
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 78
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 156
¿Quiere aparecer en la lista? [Envíe una plantilla de la comunidad](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) - Access to secretsline 313
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/cloudflare.mdx
- Access to secretsline 98
[Credential-related source omitted] - Network accessline 115
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 165
[Credential-related source omitted] - Network accessline 209
limpieza de ejecuciones interrumpidas se realizan en un barrido que se ejecuta una vez por minuto. La compilación añade un [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) - Network accessline 210
una vez por minuto. La compilación añade un [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) `* * * * *` al `wrangler.json` generado y conserva los Cron Triggers que ya - Network accessline 211
[Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/)
packages/core/docs/content/locales/es-ES/code-agents-ui.mdx
- Network accessline 208
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 209
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 361
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 362
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard"
packages/core/docs/content/locales/es-ES/create-with-an-agent.mdx
- Network accessline 14
- Un agente de IA que pueda ejecutar comandos de terminal, como Claude Code, Codex o Cursor. Los agentes que solo chatean, como ChatGPT o Claude en la web, también pueden ayudar. Consulta [Agentes que no pueden ejecutar comandos](#chat-only - Network accessline 15
- [Node.js](https://nodejs.org) 22.22 o posterior. - Network accessline 40
plano. 6. Lee la URL local que imprime `pnpm dev`. Comprueba que responde, por ejemplo con `curl -I <url>`. No abras un navegador ni uses - Network accessline 41
6. Lee la URL local que imprime `pnpm dev`. Comprueba que responde, por ejemplo con `curl -I <url>`. No abras un navegador ni uses herramientas de navegador headless. Yo abriré la app. - Network accessline 42
por ejemplo con `curl -I <url>`. No abras un navegador ni uses - Network accessline 61
- Una nueva carpeta de la app, llamada `my-app` salvo que la hayas cambiado, que contiene `actions/`, `app/` y un archivo `AGENTS.md`. - `pnpm dev` en ejecución sin errores. - Una URL local, normalmente `http://localhost:8080`, que el agent - Network accessline 63
- Una URL local, normalmente `http://localhost:8080`, que el agente ha comprobado que responde.
packages/core/docs/content/locales/es-ES/creating-templates.mdx
- Access to secretsline 487
[Credential-related source omitted] - Network accessline 532
[Envía una plantilla de la comunidad](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) cuando el repositorio esté listo. Estas plantillas permanecen separadas del catálogo propio y las mantienen sus auto
packages/core/docs/content/locales/es-ES/decisions.mdx
- Network accessline 18
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 19
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 20
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/locales/es-ES/deno-deploy.mdx
- Network accessline 26
También necesitas una cuenta de Deno Deploy y tener [Deno](https://docs.deno.com/runtime/getting_started/installation/) - Access to secretsline 100
[Credential-related source omitted] - Network accessline 117
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 168
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/deployment-environment-variables.mdx
- Network accessline 32
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | Omite el DDL de esquema bajo demanda en implementaciones que la omisión automática en serverless no cubre. Requiere que tu esquema ya esté migrado. Consulta [Configuración automática del esque - Network accessline 34
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | Lista de permitidos, separada por comas y solo para la devolución de llamada corporativa, de orígenes de vista previa exactos y de confianza (por ejemplo, `https://0123456789abcdef01234567--co - Network accessline 50
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 51
| `BETTER_AUTH_SECRET` | Cadena aleatoria de 32 o más caracteres. Firma las cookies de sesión y es el valor de respaldo para `OAUTH_STATE_SECRET` y los valores cifrados locales de la aplicación. También es un candidat - Network accessline 52
| `BETTER_AUTH_URL` | Anulación opcional del origen público de esta aplicación (por ejemplo, `https://mail.example.com`). El framework lo infiere del contexto de la solicitud, la plantilla conocida y la plataforma - Access to secretsline 53
[Credential-related source omitted] - Access to secretsline 63
[Credential-related source omitted] - Access to secretsline 90
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/deployment.mdx
packages/core/docs/content/locales/es-ES/dispatch.mdx
- Network accessline 116
Dispatch puede ser el único conector MCP para agentes externos: agregue `https://dispatch.agent-native.com/mcp` una vez en Claude, ChatGPT, Codex o Cursor, y una autorización llega a cada aplicación de espacio de trabajo otorgada en lugar d - Network accessline 125
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p
packages/core/docs/content/locales/es-ES/docker.mdx
- Dangerous shell operationline 70
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 72
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 113
[Credential-related source omitted] - Access to secretsline 169
[Credential-related source omitted] - Network accessline 186
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/locales/es-ES/embedding-sdk.mdx
- Network accessline 224
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 225
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 226
url="https://assets.agent-native.com/picker" - Network accessline 279
getMcpUrl("https://assets.agent-native.com"); - Network accessline 280
getA2AUrl("https://assets.agent-native.com"); - Network accessline 282
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 283
"https://assets.agent-native.com", - Network accessline 304
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 305
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 306
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 388
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 390
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 392
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 393
hostOrigin: "https://app.example.com", - Network accessline 396
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 397
hostOrigin: "https://app.example.com", - Network accessline 401
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 402
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 403
{ hostOrigin: "https://app.example.com" }, - Network accessline 407
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 408
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 409
{ hostOrigin: "https://app.example.com" }, - Network accessline 412
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 413
hostOrigin: "https://app.example.com", - Network accessline 468
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 469
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 470
hostOrigin: "https://app.example.com",
packages/core/docs/content/locales/es-ES/extensions.mdx
- Network accessline 241
| `dbQuery(sql, args)` | Leer desde SQL (con alcance automático para el usuario) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` | L - Network accessline 242
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` | A - Network accessline 243
| `extensionFetch(url, options)` | Accede a API externos a través de un proxy seguro con secretos | `extensionFetch('https://api.github.com/user')` | - Network accessline 271
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 272
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/locales/es-ES/external-agents.mdx
- Network accessline 111
https://dispatch.agent-native.com/mcp - Network accessline 122
https://mail.agent-native.com/mcp - Network accessline 126
Cada aplicación alojada también tiene una página de ayuda en `https://<app>/mcp/connect` con el URL copiable y - Network accessline 127
`https://<app>/mcp/connect` con el URL copiable y - Network accessline 234
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 278
El archivo `skills` CLI sin formato instala únicamente archivos `SKILL.md`; Clientes locales MCP todavía necesita un conector como `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 279
necesita un conector como `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 285
La selección de clientes predeterminada son todos los clientes locales compatibles; agregue `--client codex`, `--client claude-code` u otro objetivo específico para limitar la configuración. Los hosts en línea (ChatGPT, el chat de Claude en - Network accessline 291
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 306
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 309
`reconnect` encuentra cualquier entrada de configuración MCP cuyo URL termine en `/mcp` para el host dado y el cliente seleccionado (que coincida con URL independientemente del nombre del conector), luego actualiza o reemplaza el material d - Network accessline 317
Para clientes MCP que no pueden agregar un OAuth URL remoto directamente, abra la aplicación en su navegador y use su opción **Connect** (servida en `https://<app>/mcp/connect`). Una vez que haya iniciado sesión, haga clic en **Conect - Network accessline 323
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 324
[Credential-related source omitted] - Network accessline 325
"url": "https://mail.agent-native.com/mcp", - Network accessline 341
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 342
https://dispatch.agent-native.com/mcp - Network accessline 345
Esta es la misma entrada exclusiva para URL que `npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code` escribe para usted. Luego ejecute `/mcp` en el código Claude y elija **Autenticar**. El cliente d - Network accessline 348
https://dispatch.agent-native.com/mcp - Network accessline 490
npx @agent-native/core@latest connect https://plan.agent-native.com --códice del cliente --catálogo completo - Network accessline 541
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 542
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Access to secretsline 776
[Credential-related source omitted] - Access to secretsline 792
[Credential-related source omitted] - Access to secretsline 798
[Credential-related source omitted] - Network accessline 804
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 805
[Credential-related source omitted] - Network accessline 806
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 807
[Credential-related source omitted] - Network accessline 829
`connect dev` reescribe los mismos nombres de servidor MCP estables (`agent-native-mail`, `agent-native-calendar`, etc.) en la puerta de enlace de desarrollo diferido local, por lo que los nombres de las herramientas no cambian. Realiza una - Network accessline 856
- Conecte su propio agente a Dispatch con `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; use una aplicación directa URL solo cuando desee una aplicación aislada. - Access to secretsline 876
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/file-uploads.mdx
- Network accessline 134
Registra un proveedor personalizado solo para un tipo distinto de backend. Este ejemplo carga a través de una API HTTP genérica usando `fetch()`: - Network accessline 135
Este ejemplo carga a través de una API HTTP genérica usando `fetch()`: - Network accessline 152
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 153
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 154
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/locales/es-ES/getting-started.mdx
- Network accessline 23
- [Node.js](https://nodejs.org) 22.22 o posterior - Network accessline 24
- [pnpm](https://pnpm.io) instalado y disponible en tu `PATH` - Network accessline 158
Reinicia el servidor de desarrollo para que detecte la nueva ruta y abre `http://localhost:<port>/hello`. - Network accessline 159
`http://localhost:<port>/hello`.
packages/core/docs/content/locales/es-ES/harness-agents.mdx
- Network accessline 111
Agent-Native puede actuar como [ACP](https://agentclientprotocol.com) (Cliente Agente
packages/core/docs/content/locales/es-ES/key-concepts.mdx
- Network accessline 332
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 333
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 334
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/locales/es-ES/koyeb.mdx
packages/core/docs/content/locales/es-ES/mcp-apps.mdx
- Network accessline 23
Las extensiones de plugins de OpenAI añaden puntos de acceso junto a las herramientas MCP invocadas por el modelo. Cuando las aplicaciones MCP en línea están habilitadas, la herramienta compartida `open_app` anuncia un punto de acceso en la - Network accessline 261
Los hosts que muestran MCP Apps en línea incluyen el chat de Claude en la web, escritorio y móvil, además de la pestaña [Code](https://github.com/anthropics/claude-code/issues/103809) de Claude Desktop; ChatGPT; VS Code GitHub Copilot (1.10 - Network accessline 263
Utilice el enlace profundo de respaldo en la terminal de Claude Code, Cowork, Code en claude.ai y la CLI de Codex. [Cowork aún no declara compatibilidad con MCP Apps y, si se usa con el mismo conector, las aplicaciones en línea pueden desap - Access to secretsline 269
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/mcp-clients.mdx
- Network accessline 149
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 150
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 179
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 180
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 181
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 227
[Credential-related source omitted] - Network accessline 276
El ajuste preestablecido de Atlassian usa el punto de conexión OAuth 2.1 de Atlassian Rovo MCP, `https://mcp.atlassian.com/v1/mcp/authv2`, y proporciona herramientas de Jira y Confluence. Para crear incidencias de Jira se necesitan los perm - Network accessline 277
`https://mcp.atlassian.com/v1/mcp/authv2`, y proporciona herramientas de Jira y Confluence. Para crear incidencias de Jira se necesitan los permisos de Atlassian que permiten escribir en Jira; el usuario solo puede ver los proyectos y el co - Network accessline 279
Los administradores de una organización de Atlassian pueden controlar si un cliente OAuth tiene permiso para conectarse. Antes de establecer la conexión, es posible que un administrador deba añadir el origen HTTPS real de la aplicación a la - Network accessline 285
1. **Nombre del servidor**: una etiqueta breve para su propia referencia (por ejemplo, "zapier", "slack-tools"). 2. **URL**: el punto final HTTPS que le proporcionó el proveedor del servidor MCP (por ejemplo, `https://mcp.zapier.com/s/abc12 - Network accessline 286
2. **URL**: el punto final HTTPS que le proporcionó el proveedor del servidor MCP (por ejemplo, `https://mcp.zapier.com/s/abc123/mcp`). Esto generalmente se encuentra en el panel del proveedor o en los documentos de integración. - Network accessline 301
[Credential-related source omitted] - Network accessline 394
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 395
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 396
"url": "https://example.com/mcp", - Network accessline 434
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com
packages/core/docs/content/locales/es-ES/mcp-protocol.mdx
- Access to secretsline 31
[Credential-related source omitted] - Network accessline 108
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 117
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 118
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 119
"url": "https://mail.example.com/mcp", - Network accessline 138
POST https://your-app.example.com/mcp - Access to secretsline 143
[Credential-related source omitted] - Access to secretsline 265
[Credential-related source omitted] - Network accessline 273
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 279
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 280
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 395
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 404
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 405
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 406
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/locales/es-ES/messaging.mdx
- Network accessline 81
1. Vaya a **[api.slack.com/apps](https://api.slack.com/apps)** y haga clic en **Crear nueva aplicación** → **Desde cero**. Asígnale un nombre (por ejemplo, "Agente") y elige tu espacio de trabajo. - Network accessline 97
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 120
https://your-clips.example.com/api/slack/unfurl - Access to secretsline 134
[Credential-related source omitted] - Network accessline 169
Configure OAuth Redirect URL como `https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 170
`https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 171
Event Subscriptions Request URL como `https://your-app.example.com/_agent-native/integrations/slack/webhook` e Interactivity Request URL como - Network accessline 172
`https://your-app.example.com/_agent-native/integrations/slack/webhook` e - Network accessline 173
Interactivity Request URL como `https://your-app.example.com/_agent-native/integrations/slack/interactions`. Suscríbase a `app_home_opened`, `app_context_changed`, `app_mention`, - Network accessline 174
`https://your-app.example.com/_agent-native/integrations/slack/interactions`. - Access to secretsline 177
[Credential-related source omitted] - Network accessline 273
[Credential-related source omitted] - Access to secretsline 273
[Credential-related source omitted] - Access to secretsline 278
[Credential-related source omitted] - Access to secretsline 375
[Credential-related source omitted] - Network accessline 389
1. Abre Telegram y envía un mensaje **[@BotFather](https://t.me/BotFather)**. - Network accessline 396
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 415
1. Regístrese en **[resend.com](https://resend.com)**. El nivel gratuito es suficiente para empezar. - Network accessline 422
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 429
[Credential-related source omitted] - Network accessline 436
1. Regístrese en **[sendgrid.com](https://sendgrid.com)**. - Network accessline 444
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 449
[Credential-related source omitted] - Network accessline 473
1. Vaya a **[Meta Developer Portal](https://developers.facebook.com/)**, haga clic en **Crear aplicación** y elija el tipo **Empresa**. - Access to secretsline 480
[Credential-related source omitted] - Network accessline 486
Devolución de llamada URL: https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Access to secretsline 697
[Credential-related source omitted] - Access to secretsline 698
[Credential-related source omitted] - Access to secretsline 808
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/multi-app-workspace.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 100
[Credential-related source omitted] - Network accessline 245
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 246
"urls": { "design": "https://design.example.test" } - Network accessline 247
"design": "https://design.example.test" - Access to secretsline 276
[Credential-related source omitted] - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Network accessline 325
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 326
# https://your-agents.com/mail/* → apps/mail - Network accessline 327
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 328
# https://your-agents.com/forms/* → apps/forms
packages/core/docs/content/locales/es-ES/native-chat-ui.mdx
- Network accessline 453
Consulta la [guía de migración del chat de AgentKit](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md) para ver el mapa de API anterior y nuevo y los pasos de migración.
packages/core/docs/content/locales/es-ES/netlify.mdx
packages/core/docs/content/locales/es-ES/node-js.mdx
packages/core/docs/content/locales/es-ES/notifications.mdx
- Network accessline 192
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 193
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 194
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 211
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 212
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 213
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/locales/es-ES/observability.mdx
- Network accessline 509
El directorio [`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel) del repositorio contiene una versión más completa de este plugin, con endpoints por señal y validación del exportador. No está publicado en n
packages/core/docs/content/locales/es-ES/onboarding.mdx
- Access to secretsline 187
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/plan-plugin.mdx
- Network accessline 14
- **Tres skills** — `/visual-plan` (el punto de entrada canónico), `/visual-recap` y `/visualize-repo`. - **El conector MCP de Plan** — registrado contra la app alojada en `https://plan.agent-native.com` (endpoint MCP `https://plan.agent-na - Network accessline 15
- **El conector MCP de Plan** — registrado contra la app alojada en `https://plan.agent-native.com` (endpoint MCP `https://plan.agent-native.com/mcp`, nombre de servidor `plan`). - Network accessline 91
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 116
- **Usa un navegador Chromium.** Safari bloquea que la página HTTPS alojada de Plan lea el puente localhost `http://127.0.0.1` (contenido mixto / red privada), por lo que la página se - Network accessline 117
puente localhost `http://127.0.0.1` (contenido mixto / red privada), por lo que la página se - Network accessline 208
- `--client codex|claude-code|claude-code-cli|cowork|all` — para qué agentes locales escribir la configuración MCP (por defecto `all`). - `--no-connect` — registra el conector sin autenticar; ejecuta más tarde `npx @agent-native/core@latest - Network accessline 209
- `--no-connect` — registra el conector sin autenticar; ejecuta más tarde `npx @agent-native/core@latest connect https://plan.agent-native.com --client all`, o elige un `--client` más específico. - Network accessline 250
Tras la instalación, **inicia un nuevo hilo de Codex** para que las skills y las herramientas MCP se carguen en la sesión. El plugin incluye un conector solo con URL (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`); `codex mcp l - Network accessline 254
**Instalaciones antiguas:** si tu configuración todavía tiene una entrada `agent-native-plans` que apunta a la misma URL, ejecutar `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` para Codex, o el mi
packages/core/docs/content/locales/es-ES/pr-visual-recap.mdx
- Access to secretsline 196
[Credential-related source omitted] - Access to secretsline 197
[Credential-related source omitted] - Access to secretsline 198
[Credential-related source omitted] - Network accessline 213
[Credential-related source omitted] - Access to secretsline 213
[Credential-related source omitted] - Access to secretsline 228
[Credential-related source omitted] - Access to secretsline 243
[Credential-related source omitted] - Network accessline 256
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 270
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 283
[Credential-related source omitted] - Network accessline 289
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | Variable, solo para la variante de copia (el llamador reutilizable la ignora). Establécela en `false` para dejar de generar el reca - Network accessline 291
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | Secret. Solo cuando autoalojas la app Plans en un origen diferente. - Network accessline 334
[Credential-related source omitted] - Access to secretsline 334
[Credential-related source omitted] - Instruction override signalline 429
- Líneas que se leen como comandos directos o instrucciones de rol ("Ignore previous instructions...", "You are now...", "Write the token to..."). - Access to secretsline 541
[Credential-related source omitted] - Access to secretsline 596
[Credential-related source omitted] - Access to secretsline 597
[Credential-related source omitted] - Access to secretsline 598
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/privacy-and-data-rights.mdx
- Network accessline 59
- [Texto del GDPR](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [Comisión Europea: solicitudes individuales](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individ - Network accessline 60
- [Comisión Europea: solicitudes individuales](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [Fiscalía General de California: CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [Reglamentos de CCPA](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/locales/es-ES/real-time-collaboration.mdx
- Network accessline 687
`GET /_agent-native/health` informa del resultado, y resolverlo realiza el registro, así que un solo curl activa y confirma a la vez: - Network accessline 688
registro, así que un solo curl activa y confirma a la vez: - Network accessline 691
curl -s https://your-app.example.com/_agent-native/health | jq .realtime
packages/core/docs/content/locales/es-ES/recurring-jobs.mdx
packages/core/docs/content/locales/es-ES/render.mdx
packages/core/docs/content/locales/es-ES/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- Nunca utilice `innerHTML`, `eval()` o `document.write()` - Access to secretsline 395
[Credential-related source omitted] - Access to secretsline 411
[Credential-related source omitted] - Access to secretsline 417
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/server-database.mdx
- Network accessline 8
Toda aplicación nativa del agente almacena su estado en PostgreSQL. La UI y el agente leen y escriben las mismas tablas a través de las mismas actions, usando el mismo cliente [Drizzle ORM](https://orm.drizzle.team). La sincronización en vi - Network accessline 81
[PGlite](https://pglite.dev/) ejecuta una base de datos Postgres WASM dentro del mismo proceso. Le permite detectar problemas de esquema exclusivos de Postgres antes de desplegar, manteniendo una configuración tan sencilla como la del valor
packages/core/docs/content/locales/es-ES/server-middleware.mdx
- Network accessline 106
Para más patrones de middleware — lógica condicional, operaciones asíncronas, interceptación de respuestas — consulte los [documentos de middleware de Nitro](https://nitro.build/guide/routing#middleware) y los [documentos de H3](https://h3. - Network accessline 149
- [**Rutas**](/docs/server-routes): rutas personalizadas protegidas por este middleware - [**Seguridad**](/docs/security): delimitación de datos, guardias de acceso y el modelo completo de autenticación - [**Documentos de middleware de Nitr - Network accessline 151
- [**Documentos de middleware de Nitro**](https://nitro.build/guide/routing#middleware): la API completa de middleware y sus patrones
packages/core/docs/content/locales/es-ES/server-overview.mdx
- Network accessline 206
[Nitro](https://nitro.build) es el conjunto de herramientas de servidor que impulsa la capa HTTP nativa del agente. La mayor parte de lo que configurará en `server/` es Nitro: rutas basadas en archivos, plugins de inicio y middleware. - Network accessline 214
La API subyacente de solicitud/respuesta es [H3](https://h3.unjs.io). Cuando vea `defineEventHandler`, `createError` o `getQuery` en rutas personalizadas, eso es H3. Los [documentos de Nitro](https://nitro.build/guide) y los [documentos de
packages/core/docs/content/locales/es-ES/server-plugins.mdx
- Network accessline 141
Los plugins pueden usar cualquiera de los ganchos del ciclo de vida de Nitro: `request`, `beforeResponse`, `close` y otros. Los [documentos de plugins de Nitro](https://nitro.build/guide/plugins) cubren en detalle la API completa de ganchos - Network accessline 163
- [**Rutas**](/docs/server-routes): rutas de archivo personalizadas y cuándo usarlas - [**Redacción de Instrucciones para el Agente**](/docs/writing-agent-instructions): cómo redactar el `systemPrompt` en el plugin de agent-chat - [**Docume - Network accessline 165
- [**Documentos de plugins de Nitro**](https://nitro.build/guide/plugins): la API completa de plugins y sus ganchos de ciclo de vida
packages/core/docs/content/locales/es-ES/server-routes.mdx
- Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/skills-guide.mdx
- Network accessline 315
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 316
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 317
"url": "https://assets.agent-native.com", - Network accessline 318
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/locales/es-ES/template-analytics.mdx
packages/core/docs/content/locales/es-ES/template-assets.mdx
- Network accessline 117
Demostración en vivo: [assets.agent-native.com](https://assets.agent-native.com). - Network accessline 466
<EmbeddedApp url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 467
url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 503
La habilidad de la aplicación Assets tiene el ID de aplicación `assets` y está alojado MCP URL `https://assets.agent-native.com/mcp`. - Network accessline 504
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/locales/es-ES/template-brain.mdx
- Network accessline 159
Demostración en vivo: [brain.agent-native.com](https://brain.agent-native.com).
packages/core/docs/content/locales/es-ES/template-calendar-developers.mdx
- Network accessline 30
Esto necesita [Node.js](https://nodejs.org) 22.22 o posterior y - Network accessline 31
[pnpm](https://pnpm.io) instalados primero. Consulta [Primeros - Network accessline 42
La app se abre en `http://localhost:8082`. Si no se abre automáticamente, - Network accessline 53
1. **Crea un proyecto de Google Cloud**, si aún no tienes uno para usar: en [Google Cloud Console](https://console.cloud.google.com/), haz clic - Network accessline 54
en [Google Cloud Console](https://console.cloud.google.com/), haz clic - Network accessline 65
4. **Crea las credenciales**: en **APIs & Services → Credentials**, crea un ID de cliente OAuth de tipo "Web application," y añade `http://localhost:8082/_agent-native/google/callback` como URI de - Network accessline 66
un ID de cliente OAuth de tipo "Web application," y añade `http://localhost:8082/_agent-native/google/callback` como URI de redirección autorizada (8082 es el puerto de desarrollo predeterminado - Network accessline 67
`http://localhost:8082/_agent-native/google/callback` como URI de - Access to secretsline 85
[Credential-related source omitted] - Network accessline 88
Obtén un Client ID y Secret desde una [app OAuth de Zoom](https://marketplace.zoom.us/) (Zoom Marketplace → Develop → Build App - Network accessline 89
Zoom](https://marketplace.zoom.us/) (Zoom Marketplace → Develop → Build App
packages/core/docs/content/locales/es-ES/template-calendar-features.mdx
packages/core/docs/content/locales/es-ES/template-calendar-integrations.mdx
- Access to secretsline 24
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/template-calendar.mdx
packages/core/docs/content/locales/es-ES/template-chat-developers.mdx
- Network accessline 31
Esto necesita [Node.js](https://nodejs.org) 22.22 o posterior. Si `pnpm` no
packages/core/docs/content/locales/es-ES/template-chat-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/template-chat.mdx
packages/core/docs/content/locales/es-ES/template-clips-developers.mdx
- Network accessline 22
Esto necesita tener instalados antes [Node.js](https://nodejs.org) 22.22 o posterior y [pnpm](https://pnpm.io). Consulta [Primeros pasos](/docs/getting-started) si todavía no has instalado ninguno de los dos, o si no tienes configurada una - Network accessline 305
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 306
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 307
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/template-clips-embed.mdx
- Network accessline 40
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 53
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 54
[Credential-related source omitted] - Network accessline 56
const response = await fetch( - Network accessline 145
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 152
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/locales/es-ES/template-clips-features.mdx
- Network accessline 117
La aplicación de escritorio está diseñada para la captura diaria más fluida: grabación de pantalla y cámara web con una burbuja de cámara que puedes arrastrar, transcripción en vivo mientras grabas, todo sin mantener abierta una pestaña del - Network accessline 212
[Credential-related source omitted] - Network accessline 213
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/template-clips-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/template-clips.mdx
packages/core/docs/content/locales/es-ES/template-content.mdx
- Network accessline 17
### [Pruébalo ahora →](https://content.agent-native.com)
packages/core/docs/content/locales/es-ES/template-design-developers.mdx
packages/core/docs/content/locales/es-ES/template-design-integrations.mdx
- Access to secretsline 25
[Credential-related source omitted] - Network accessline 51
https://design.agent-native.com/mcp - Network accessline 56
ejecuta `/mcp` para autenticarse. Codex y otros clientes CLI o stdio en cambio ejecutan `npx @agent-native/core@latest connect https://design.agent-native.com`, lo que abre una pestaña del navegador para - Network accessline 57
cambio ejecutan `npx @agent-native/core@latest connect https://design.agent-native.com`, lo que abre una pestaña del navegador para autorizar y genera un token por usuario, revocable, sin ningún secreto - Network accessline 58
https://design.agent-native.com`, lo que abre una pestaña del navegador para - Network accessline 74
devolver un error de sesión o autenticación, no sigas reintentándola. Ejecuta `/mcp` en Claude Code y vuelve a conectar, o ejecuta `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` desde - Network accessline 75
Ejecuta `/mcp` en Claude Code y vuelve a conectar, o ejecuta `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` desde cualquier terminal. Ambos vuelven a autenticarse sin reinstalar nada. - Network accessline 76
@agent-native/core@latest reconnect https://design.agent-native.com` desde
packages/core/docs/content/locales/es-ES/template-design.mdx
packages/core/docs/content/locales/es-ES/template-dispatch.mdx
packages/core/docs/content/locales/es-ES/template-forms-agent.mdx
- Network accessline 40
- **"Publica esto y dame el enlace."** El formulario se activa y obtienes su URL pública de inmediato. - **"Crea un formulario anónimo de comentarios de clientes y dame el enlace público."** El formulario se publica anónimo desde el inicio, - Network accessline 42
- **"Solo permite que este formulario se envíe desde https://example.com."** El formulario empieza a rechazar envíos desde cualquier otro lugar. Esto solo está disponible pidiéndoselo al agente hoy en día. No hay un interruptor para ello en
packages/core/docs/content/locales/es-ES/template-forms-features.mdx
- Network accessline 14
[Credential-related source omitted] - Network accessline 15
[Credential-related source omitted] - Network accessline 94
Los envíos públicos pueden requerir verificación mediante [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/) antes de pasar. Actívalo definiendo una clave de sitio y una clave secreta en Configuración. Dos protecciones s - Network accessline 110
Hoy no hay un interruptor para esto en el creador. Pídele al agente que lo configure, por ejemplo "solo permite que este formulario se envíe desde https://example.com." Dale al agente el origen exacto (el esquema y el dominio, como `https:/ - Network accessline 116
[Credential-related source omitted] - Network accessline 117
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/template-forms.mdx
packages/core/docs/content/locales/es-ES/template-mail.mdx
- Network accessline 97
Demostración en vivo: [mail.agent-native.com](https://mail.agent-native.com). - Network accessline 168
1. Abre [Google Cloud Console](https://console.cloud.google.com/) y crea un proyecto. - Network accessline 169
2. Habilite **Gmail API** en APIs & Services → Biblioteca. 3. Cree credenciales OAuth 2.0 (tipo: aplicación web). Añade `http://localhost:8085/_agent-native/google/callback` como redireccionamiento autorizado URI. 4. Copie el ID del cliente - Network accessline 170
3. Cree credenciales OAuth 2.0 (tipo: aplicación web). Añade `http://localhost:8085/_agent-native/google/callback` como redireccionamiento autorizado URI.
packages/core/docs/content/locales/es-ES/template-plan.mdx
- Network accessline 143
y `/visualize-repo` — además del conector MCP `plan` (un servidor de herramientas remoto con el que habla tu agente) en `https://plan.agent-native.com/mcp`. Para la referencia completa de flags, las - Network accessline 144
herramientas remoto con el que habla tu agente) en `https://plan.agent-native.com/mcp`. Para la referencia completa de flags, las rutas de instalación por plugin/marketplace para Claude Code y Codex, y la - Network accessline 145
`https://plan.agent-native.com/mcp`. Para la referencia completa de flags, las - Network accessline 244
`Unauthorized` o `Session terminated`, no sigas reintentándolo. Vuelve a autenticarte con `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, - Network accessline 245
autenticarte con `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, o vuelve a ejecutar `/mcp` → **Authenticate** para un host compatible con - Network accessline 246
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`,
packages/core/docs/content/locales/es-ES/template-slides-features.mdx
packages/core/docs/content/locales/es-ES/template-slides-integrations.mdx
- Access to secretsline 25
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/template-slides.mdx
packages/core/docs/content/locales/es-ES/toolkit-capability-packages.mdx
packages/core/docs/content/locales/es-ES/toolkit-settings.mdx
- Access to secretsline 121
[Credential-related source omitted] - Network accessline 191
El flujo de autoservicio de OpenAI requiere una devolución de llamada OAuth HTTP en `127.0.0.1`, una dirección de bucle local en el equipo del usuario. Por eso no se admiten las aplicaciones alojadas, incluidos los sitios alojados de `*.age - Network accessline 201
Usa solo la Responses API pública para la inferencia. Los endpoints del backend privado de ChatGPT y los datos de sus conversaciones no forman parte de esta autorización. OpenAI marca el flujo de uso compartido de tokens como una vista prev
packages/core/docs/content/locales/es-ES/tracking.mdx
- Network accessline 83
| Proveedor | Variaciones de entorno | | ----------- | --------------------------------------------------------------------------------------------------- | | Co - Network accessline 84
| ----------- | --------------------------------------------------------------------------------------------------- | | Correo | `POSTHOG_API_KEY` (obligatorio), `POSTHOG_HOST` (opcional, por defecto `https://us.i.posthog.com`) | | Pa - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Network accessline 205
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 206
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 207
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 225
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 226
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 283
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 284
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 285
fetch("https://analytics.example.com/events", { - Network accessline 343
- **HTTP por lotes**: los proveedores integrados ponen en cola los eventos y los vacían cada 10 segundos o cuando se acumulan 50 eventos, lo que ocurra primero. Esto minimiza las solicitudes salientes sin perder datos. - **Sin dependencias - Network accessline 344
- **Sin dependencias de SDK**: todos los proveedores integrados utilizan `fetch()` sin formato. Sin PostHog SDK, sin Mixpanel SDK, sin Amplitude SDK. Mantiene el marco liviano.
packages/core/docs/content/locales/es-ES/turn-into-app.mdx
- Network accessline 12
- Un agente de IA que pueda ejecutar comandos de terminal, como Claude Code, Codex o Cursor. Claude o ChatGPT en la web pueden entregarle el trabajo a Builder en su lugar. Consulta [Dónde se ejecuta](#hosts). - [Node.js](https://nodejs.org) - Network accessline 13
- [Node.js](https://nodejs.org). El instalador de nodejs.org incluye `npx`, que ejecuta los comandos de instalación de más abajo. - Network accessline 18
- **En tu equipo.** Node.js 22.22 o posterior (compruébalo con `node --version`) y [pnpm](https://pnpm.io) en tu `PATH`. Si `pnpm --version` falla, ejecuta `corepack enable`, que viene incluido con Node.js. - Network accessline 59
El alojamiento en Builder, desde cualquiera de los dos tipos de host, requiere el conector MCP de Agent-Native Dispatch (`https://dispatch.agent-native.com/mcp`, OAuth) y llama a su acción `start-workspace-app-creation`. En tu lado no se ej - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/updating-ui-in-production.mdx
- Network accessline 20
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment) es un flujo de trabajo gestionado independiente con capacidad de código donde un agente puede modificar la UI de tu a
packages/core/docs/content/locales/es-ES/upgrading-core-ui.mdx
- Access to secretsline 73
[Credential-related source omitted] - Access to secretsline 76
[Credential-related source omitted] - Network accessline 189
URL canónica de la fuente: [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx). - Network accessline 190
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx).
packages/core/docs/content/locales/es-ES/vercel.mdx
- Access to secretsline 94
[Credential-related source omitted] - Network accessline 111
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 138
[Credential-related source omitted] - Network accessline 202
limpieza de ejecuciones interrumpidas se realizan en un barrido que se ejecuta una vez por minuto. La compilación añade un [trabajo cron de Vercel](https://vercel.com/docs/cron-jobs) que llama al barrido - Network accessline 203
una vez por minuto. La compilación añade un [trabajo cron de Vercel](https://vercel.com/docs/cron-jobs) que llama al barrido en `/_agent-native/jobs/_process-sweep`. Una compilación de espacio de trabajo - Network accessline 204
[trabajo cron de Vercel](https://vercel.com/docs/cron-jobs) que llama al barrido
packages/core/docs/content/locales/es-ES/voice-input.mdx
- Access to secretsline 14
[Credential-related source omitted] - Possible obfuscationline 24
3. **Voz web del navegador API (alternativa).** Si no hay ningún proveedor de servidor disponible, el compositor puede utilizar el reconocimiento de voz integrado del navegador. Funciona en navegadores basados <U+200B><U+200B>en Chromium (Chrome, Edge, A - Access to secretsline 88
[Credential-related source omitted]
packages/core/docs/content/locales/es-ES/what-is-agent-native.mdx
packages/core/docs/content/locales/es-ES/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Network accessline 271
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 272
[Credential-related source omitted] - Network accessline 273
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 274
[Credential-related source omitted] - Access to secretsline 275
[Credential-related source omitted] - Access to secretsline 276
[Credential-related source omitted] - Network accessline 288
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 289
[Credential-related source omitted] - Network accessline 290
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 291
[Credential-related source omitted] - Access to secretsline 292
[Credential-related source omitted] - Access to secretsline 293
[Credential-related source omitted] - Network accessline 345
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 347
APP_URL=https://workspace.example.com
packages/core/docs/content/locales/es-ES/workspace-management.mdx
- Network accessline 90
No todos los que necesitan realizar cambios se sienten cómodos con git. [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management) admite un modelo de ramificación vi - Network accessline 140
Consejos clave: utilice equipos GitHub (`@org/team`), no individuos. Los cambios en el marco y el envío siempre deben requerir una revisión de la plataforma. Consulte [GitHub CODEOWNERS docs](https://docs.github.com/en/repositories/managing - Network accessline 160
Luego agregue la acción [actions/labeler](https://github.com/actions/labeler); consulte el README de ese repositorio para conocer el flujo de trabajo completo YAML. Las etiquetas se aplican automáticamente cuando se abren o actualizan los P - Access to secretsline 233
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/a2a-protocol.mdx
- Access to secretsline 93
[Credential-related source omitted] - Network accessline 117
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 118
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 119
"url": "https://analytics.example.com", - Access to secretsline 162
[Credential-related source omitted] - Network accessline 314
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 350
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 351
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 352
"https://analytics.example.com", - Access to secretsline 354
[Credential-related source omitted] - Network accessline 374
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", - Network accessline 375
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", "cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 376
"url": "https://example.services.ai.azure.com/a2a", - Network accessline 377
"cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 417
[Les agents hébergés Azure AI Foundry](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 424
Foundry peut exposer un point de terminaison A2A via [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint). - Network accessline 434
[Gemini Enterprise peut enregistrer des agents A2A externes](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent), - Network accessline 437
`GET {url}/v1/card`, l'appel `POST {url}/v1/message:send` et le streaming `POST {url}/v1/message:stream`. Les requêtes utilisent des jetons Bearer Google OAuth ou ADC et les autorisations Google Cloud IAM. Consultez [Appeler un agent avec s - Network accessline 438
`POST {url}/v1/message:stream`. Les requêtes utilisent des jetons Bearer Google OAuth ou ADC et les autorisations Google Cloud IAM. Consultez [Appeler un agent avec son point de terminaison A2A de registre](https://docs.cloud.google.com/gem - Network accessline 439
Google OAuth ou ADC et les autorisations Google Cloud IAM. Consultez [Appeler un agent avec son point de terminaison A2A de registre](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) - Network accessline 457
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 458
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 459
"url": "https://api.anthropic.com", - Access to secretsline 464
[Credential-related source omitted] - Access to secretsline 477
[Credential-related source omitted] - Network accessline 523
# Appeler une application déployée directement par URL, en lisant le jeton au porteur depuis l'environnement agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 524
[Credential-related source omitted] - Access to secretsline 524
[Credential-related source omitted] - Access to secretsline 525
[Credential-related source omitted] - Access to secretsline 599
[Credential-related source omitted] - Access to secretsline 603
[Credential-related source omitted] - Network accessline 661
- Un `A2A_SECRET` partagé est monté dans l'environnement de chaque application au moment de la construction. - Les appels entre applications ont la même origine : `https://workspace.example.com/apps/analytics` appelle `https://workspace.exa - Network accessline 662
- Les appels entre applications ont la même origine : `https://workspace.example.com/apps/analytics` appelle `https://workspace.example.com/apps/mail` – il n'y a donc pas de configuration DNS, CORS ou JWT par paire. - Network accessline 670
**Ne comptez jamais sur un `Promise` qui survivra à la réponse.** Les fonctions sans serveur (Netlify, Vercel, AWS Lambda, Cloud Run) gèlent au moment où le corps de la réponse est vidé - parfois avant même que la poignée de main TCP d'un ` - Network accessline 708
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 709
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 710
"https://analytics.example.com", - Access to secretsline 712
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/actions-defining.mdx
- Network accessline 226
`schema` accepte toute bibliothèque compatible [Standard Schema](https://standardschema.dev) :
packages/core/docs/content/locales/fr-FR/add-ui-to-your-agent.mdx
- Network accessline 27
Collez l’URL de carte A2A v1.0 et choisissez les identifiants Entra. Pour le service à service, utilisez `https://ai.azure.com/.default` et le rôle **Foundry Agent Consumer**. - Network accessline 45
Parcourez la [Global A2A Registry](https://www.a2a-registry.org). Vérifiez la carte, l’hôte et l’authentification avant d’ajouter l’agent et d’utiliser un secret.
packages/core/docs/content/locales/fr-FR/agent-mentions.mdx
packages/core/docs/content/locales/fr-FR/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/agent-web-surfaces.mdx
- Network accessline 171
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 172
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 173
siteUrl: "https://example.com", - Network accessline 198
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/locales/fr-FR/audit-log.mdx
- Network accessline 110
**`export-audit-ocsf`** renvoie le journal de l'organisation sous forme d'événements OCSF 1.9.0 API Activity ([OCSF](https://schema.ocsf.io), `class_uid` 6003), du plus ancien au plus récent, qu'un SIEM peut extraire à intervalles réguliers
packages/core/docs/content/locales/fr-FR/authentication.mdx
- Network accessline 8
Les applications natives d'agent utilisent [Better Auth](https://better-auth.com) pour l'authentification avec une conception axée sur le compte. Les utilisateurs créent un compte lors de la première visite et obtiennent une véritable ident - Access to secretsline 196
[Credential-related source omitted] - Access to secretsline 201
[Credential-related source omitted] - Access to secretsline 205
[Credential-related source omitted] - Access to secretsline 210
[Credential-related source omitted] - Access to secretsline 212
[Credential-related source omitted] - Access to secretsline 234
[Credential-related source omitted] - Access to secretsline 238
[Credential-related source omitted] - Network accessline 251
https://mail.agent-native.com/mcp - Access to secretsline 476
[Credential-related source omitted] - Access to secretsline 478
[Credential-related source omitted] - Access to secretsline 480
[Credential-related source omitted] - Access to secretsline 481
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/automations.mdx
packages/core/docs/content/locales/fr-FR/aws-amplify.mdx
- Access to secretsline 123
[Credential-related source omitted] - Network accessline 142
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 206
[Credential-related source omitted] - Network accessline 220
Créez l'application dans la [console Amplify](https://console.aws.amazon.com/amplify/) : - Network accessline 255
Le premier build prend quelques minutes, et vous pouvez le suivre dans la console Amplify. Une fois terminé, ouvrez l'adresse de l'application, qui ressemble à `https://main.d1m7bkiki6tdw1.amplifyapp.com`, et créez le premier - Network accessline 256
console Amplify. Une fois terminé, ouvrez l'adresse de l'application, qui ressemble à `https://main.d1m7bkiki6tdw1.amplifyapp.com`, et créez le premier compte. - Network accessline 257
ressemble à `https://main.d1m7bkiki6tdw1.amplifyapp.com`, et créez le premier - Access to secretsline 316
[Credential-related source omitted] - Access to secretsline 320
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/aws-lambda.mdx
- Network accessline 41
Vous avez aussi besoin d'un compte AWS et de l'[AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html), - Network accessline 42
l'[AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html), - Access to secretsline 124
[Credential-related source omitted] - Access to secretsline 162
[Credential-related source omitted] - Access to secretsline 180
[Credential-related source omitted] - Access to secretsline 230
[Credential-related source omitted] - Access to secretsline 238
[Credential-related source omitted] - Network accessline 264
La première commande affiche la `FunctionUrl`, qui ressemble à `https://abc123.lambda-url.us-east-1.on.aws/`. Copiez-la pour l'étape 7. Les - Network accessline 265
`https://abc123.lambda-url.us-east-1.on.aws/`. Copiez-la pour l'étape 7. Les - Network accessline 285
Créez la distribution CloudFront dans la [console CloudFront](https://console.aws.amazon.com/cloudfront/v4/home) : - Network accessline 286
[console CloudFront](https://console.aws.amazon.com/cloudfront/v4/home) : - Network accessline 301
Dans l'onglet **Origins** de la distribution, créez une origine. Pour **Origin domain**, collez le nom d'hôte de l'URL de fonction de l'étape 5, sans `https://` ni la barre oblique finale. Réglez **Protocol** sur **HTTPS only** et - Network accessline 302
**Origin domain**, collez le nom d'hôte de l'URL de fonction de l'étape 5, sans `https://` ni la barre oblique finale. Réglez **Protocol** sur **HTTPS only** et **Response timeout** sur 60 secondes. N'ajoutez pas de contrôle d'accès à - Network accessline 303
`https://` ni la barre oblique finale. Réglez **Protocol** sur **HTTPS only** et - Network accessline 340
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 342
[Credential-related source omitted] - Access to secretsline 344
[Credential-related source omitted] - Access to secretsline 362
[Credential-related source omitted] - Dangerous shell operationline 364
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/locales/fr-FR/azure-static-web-apps.mdx
packages/core/docs/content/locales/fr-FR/blueprint-installer.mdx
- Network accessline 68
agent-native add <kind> <name> # print a curated blueprint agent-native add <kind> <https://docs…> # rechercher et intégrer à partir d'un URL - Network accessline 69
agent-native add <kind> <https://docs…> # rechercher et intégrer à partir d'un URL - Network accessline 95
agent-native add provider https://docs.example.com/api | claude
packages/core/docs/content/locales/fr-FR/cli-adapters.mdx
- Access to secretsline 99
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/locales/fr-FR/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">Composant</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 126
optional: true, description: "Tout ce qui vient des [options de useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) sauf `queryKey` et `queryFn`, que le hook définit lui-même.", - Network accessline 127
description: "Tout ce qui vient des [options de useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) sauf `queryKey` et `queryFn`, que le hook définit lui-même.", }, - Network accessline 128
"Tout ce qui vient des [options de useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) sauf `queryKey` et `queryFn`, que le hook définit lui-même.", - Network accessline 138
Ce hook est destiné à être utilisé pour les actions en lecture seule. C'est un wrapper léger autour de la fonction [useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery). - Network accessline 174
optional: true, description: "Tout ce qui vient des [options de useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) sauf `mutationFn`, que le hook définit lui-même.", - Network accessline 175
description: "Tout ce qui vient des [options de useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) sauf `mutationFn`, que le hook définit lui-même.", fields: [ - Network accessline 176
"Tout ce qui vient des [options de useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) sauf `mutationFn`, que le hook définit lui-même.", - Network accessline 203
Ce hook est destiné à être utilisé pour les actions qui modifient des données, comme la création, la mise à jour ou l'envoi de quelque chose. C'est un wrapper léger autour de la fonction [useMutation()](https://tanstack.com/query/latest/doc
packages/core/docs/content/locales/fr-FR/cloneable-saas.mdx
- Network accessline 49
https://github.com/owner/repo - Network accessline 67
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 68
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 77
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 78
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 157
Vous souhaitez figurer dans la liste ? [Proposez un modèle communautaire](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) - Access to secretsline 314
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/cloudflare.mdx
- Access to secretsline 107
[Credential-related source omitted] - Network accessline 126
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 178
[Credential-related source omitted] - Network accessline 223
d'attente et le nettoyage des exécutions interrompues passent tous par un balayage exécuté une fois par minute. Le build ajoute un [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) - Network accessline 224
balayage exécuté une fois par minute. Le build ajoute un [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) `* * * * *` au `wrangler.json` généré et conserve les Cron Triggers que vous - Network accessline 225
[Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/)
packages/core/docs/content/locales/fr-FR/code-agents-ui.mdx
- Network accessline 207
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 208
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 360
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 361
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard"
packages/core/docs/content/locales/fr-FR/create-with-an-agent.mdx
- Network accessline 14
- Un agent IA capable d’exécuter des commandes de terminal, comme Claude Code, Codex ou Cursor. Les agents qui ne font que discuter, comme ChatGPT ou Claude sur le web, peuvent quand même aider. Voir [Agents qui ne peuvent pas exécuter de c - Network accessline 15
- [Node.js](https://nodejs.org) 22.22 ou version ultérieure. - Network accessline 41
arrière-plan. 6. Lis l’URL locale affichée par `pnpm dev`. Vérifie qu’elle répond, par exemple avec `curl -I <url>`. N’ouvre pas de navigateur et - Network accessline 42
6. Lis l’URL locale affichée par `pnpm dev`. Vérifie qu’elle répond, par exemple avec `curl -I <url>`. N’ouvre pas de navigateur et n’utilise pas d’outils de navigateur headless. J’ouvrirai - Network accessline 43
par exemple avec `curl -I <url>`. N’ouvre pas de navigateur et - Network accessline 63
- Un nouveau dossier d’application, nommé `my-app` sauf si vous l’avez changé, qui contient `actions/`, `app/` et un fichier `AGENTS.md`. - `pnpm dev` en cours d’exécution sans erreur. - Une URL locale, généralement `http://localhost:8080`, - Network accessline 65
- Une URL locale, généralement `http://localhost:8080`, dont l’agent a vérifié qu’elle répond.
packages/core/docs/content/locales/fr-FR/creating-templates.mdx
- Access to secretsline 485
[Credential-related source omitted] - Network accessline 530
[Proposez un modèle communautaire](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) lorsque le dépôt est prêt. Ces modèles restent séparés du catalogue officiel et sont maintenus par leurs auteurs.
packages/core/docs/content/locales/fr-FR/decisions.mdx
- Network accessline 18
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 19
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 20
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/locales/fr-FR/deno-deploy.mdx
- Network accessline 27
Vous avez aussi besoin d'un compte Deno Deploy et de [Deno](https://docs.deno.com/runtime/getting_started/installation/) installé sur - Network accessline 28
[Deno](https://docs.deno.com/runtime/getting_started/installation/) installé sur - Access to secretsline 108
[Credential-related source omitted] - Network accessline 127
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 182
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/deployment-environment-variables.mdx
- Network accessline 31
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | Saute le DDL de schéma à la demande sur les déploiements que le saut serverless automatique ne couvre pas. Nécessite que votre schéma soit déjà migré. Voir [DDL de schéma à la demande](/docs/d - Network accessline 33
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | Liste blanche uniquement pour le rappel corporate, séparée par des virgules, d'origines de preview de confiance exactes (par exemple, `https://0123456789abcdef01234567--content.netlify.app`). - Network accessline 47
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 48
| `BETTER_AUTH_SECRET` | Chaîne aléatoire de 32+ caractères. Signe les cookies de session et sert de repli pour `OAUTH_STATE_SECRET` et les valeurs chiffrées locales à l'application. C'est aussi un candidat de lecture - Network accessline 49
| `BETTER_AUTH_URL` | Surcharge optionnelle de l'origine publique pour cette application (par ex. `https://mail.example.com`). Le framework la déduit du contexte de la requête, du template connu, et de la plateform - Access to secretsline 50
[Credential-related source omitted] - Access to secretsline 51
[Credential-related source omitted] - Access to secretsline 59
[Credential-related source omitted] - Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/deployment.mdx
packages/core/docs/content/locales/fr-FR/dispatch.mdx
- Network accessline 116
Dispatch peut être le connecteur MCP unique pour les agents externes : ajoutez `https://dispatch.agent-native.com/mcp` une fois dans Claude, ChatGPT, Codex ou Cursor, et une autorisation atteint chaque application d'espace de travail accord - Network accessline 125
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p
packages/core/docs/content/locales/fr-FR/docker.mdx
- Dangerous shell operationline 76
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 78
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 119
[Credential-related source omitted] - Access to secretsline 179
[Credential-related source omitted] - Network accessline 199
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/locales/fr-FR/embedding-sdk.mdx
- Network accessline 224
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 225
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 226
url="https://assets.agent-native.com/picker" - Network accessline 279
getMcpUrl("https://assets.agent-native.com"); - Network accessline 280
getA2AUrl("https://assets.agent-native.com"); - Network accessline 282
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 283
"https://assets.agent-native.com", - Network accessline 304
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 305
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 306
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 388
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 390
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 392
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 393
hostOrigin: "https://app.example.com", - Network accessline 396
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 397
hostOrigin: "https://app.example.com", - Network accessline 401
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 402
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 403
{ hostOrigin: "https://app.example.com" }, - Network accessline 407
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 408
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 409
{ hostOrigin: "https://app.example.com" }, - Network accessline 412
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 413
hostOrigin: "https://app.example.com", - Network accessline 468
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 469
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 470
hostOrigin: "https://app.example.com",
packages/core/docs/content/locales/fr-FR/extensions.mdx
- Network accessline 242
| `dbQuery(sql, args)` | Lecture à partir de SQL (portée automatiquement sur l'utilisateur) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` - Network accessline 243
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` - Network accessline 244
| `extensionFetch(url, options)` | Atteignez les API externes via un proxy sécurisé avec des secrets | `extensionFetch('https://api.github.com/user')` | - Network accessline 272
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 273
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/locales/fr-FR/external-agents.mdx
- Network accessline 112
https://dispatch.agent-native.com/mcp - Network accessline 123
https://mail.agent-native.com/mcp - Network accessline 127
Chaque application hébergée dispose également d'une page d'aide à l'adresse `https://<app>/mcp/connect` avec le URL copiable et - Network accessline 128
`https://<app>/mcp/connect` avec le URL copiable et - Network accessline 236
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 280
Le `skills` brut CLI installe uniquement les fichiers `SKILL.md` ; les clients locaux MCP sont toujours besoin d'un connecteur tel que `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 281
besoin d'un connecteur tel que `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 287
La sélection de clients par défaut concerne tous les clients locaux pris en charge ; ajoutez `--client codex`, `--client claude-code` ou une autre cible spécifique pour affiner la configuration. Les hôtes avec affichage intégré (ChatGPT, le - Network accessline 293
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 308
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 311
`reconnect` trouve toute entrée de configuration MCP dont URL se termine par `/mcp` pour l'hôte donné et le client sélectionné (correspondant à URL quel que soit le nom du connecteur), puis actualise ou remplace le matériel d'authentificati - Network accessline 319
Pour les clients MCP qui ne peuvent pas ajouter directement un OAuth distant URL, ouvrez l'application dans votre navigateur et utilisez son offre **Connect** (servie sur `https://<app>/mcp/connect`). Une fois connecté, cliquez sur ** - Network accessline 325
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 326
[Credential-related source omitted] - Network accessline 327
"url": "https://mail.agent-native.com/mcp", - Network accessline 343
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 344
https://dispatch.agent-native.com/mcp - Network accessline 347
Il s'agit de la même entrée réservée uniquement à URL que `npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code` écrit pour vous. Exécutez ensuite `/mcp` dans le code Claude et choisissez **Authentifi - Network accessline 350
https://dispatch.agent-native.com/mcp - Network accessline 492
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex --full-catalog - Network accessline 543
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 544
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Access to secretsline 767
[Credential-related source omitted] - Access to secretsline 783
[Credential-related source omitted] - Access to secretsline 789
[Credential-related source omitted] - Network accessline 795
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 796
[Credential-related source omitted] - Network accessline 797
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 798
[Credential-related source omitted] - Network accessline 820
`connect dev` réécrit les mêmes noms de serveur MCP stables (`agent-native-mail`, `agent-native-calendar`, etc.) sur la passerelle de développement paresseux locale, de sorte que les noms des outils ne changent pas. Il sauvegarde les entrée - Network accessline 847
- Connectez votre propre agent à Dispatch avec `npx @agent-native/core@latest connect https://dispatch.agent-native.com` ; utilisez une application directe URL uniquement lorsque vous souhaitez une application isolée. - Access to secretsline 867
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/file-uploads.mdx
- Network accessline 134
N'enregistrez un fournisseur personnalisé que pour un type de backend différent. Cet exemple effectue un téléchargement via une API HTTP générique en utilisant `fetch()` : - Network accessline 136
en utilisant `fetch()` : - Network accessline 153
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 154
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 155
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/locales/fr-FR/getting-started.mdx
- Network accessline 23
- [Node.js](https://nodejs.org) 22.22 ou version ultérieure - Network accessline 24
- [pnpm](https://pnpm.io) installé et disponible dans votre `PATH` - Network accessline 157
Redémarrez le serveur de développement afin qu’il détecte la nouvelle route, puis ouvrez `http://localhost:<port>/hello`. - Network accessline 158
puis ouvrez `http://localhost:<port>/hello`.
packages/core/docs/content/locales/fr-FR/harness-agents.mdx
- Network accessline 111
Agent-Native peut agir comme un [ACP](https://agentclientprotocol.com) (Agent Client
packages/core/docs/content/locales/fr-FR/key-concepts.mdx
- Network accessline 331
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 332
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 333
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/locales/fr-FR/koyeb.mdx
packages/core/docs/content/locales/fr-FR/mcp-apps.mdx
- Network accessline 23
Les extensions de plugins OpenAI ajoutent des points de lancement aux outils MCP appelés par le modèle. Lorsque les applications MCP en ligne sont activées, l'outil partagé `open_app` annonce un point d'entrée dans la barre latérale globale - Network accessline 261
Les hôtes qui affichent les MCP Apps dans la conversation comprennent le chat Claude sur le Web, ordinateur et mobile, ainsi que l’onglet [Code](https://github.com/anthropics/claude-code/issues/103809) de Claude Desktop ; ChatGPT; VS Code G - Network accessline 263
Utilisez le lien profond de secours dans le terminal Claude Code, Cowork, Code sur claude.ai et la CLI Codex. [Cowork ne déclare pas encore la prise en charge des MCP Apps et son utilisation avec le même connecteur peut faire disparaître le - Access to secretsline 269
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/mcp-clients.mdx
- Network accessline 148
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 149
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 178
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 179
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 180
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 226
[Credential-related source omitted] - Network accessline 275
Le préréglage Atlassian utilise le point de terminaison OAuth 2.1 d'Atlassian Rovo MCP, `https://mcp.atlassian.com/v1/mcp/authv2`, et donne accès aux outils Jira et Confluence. La création de tickets Jira nécessite les autorisations Atlassi - Network accessline 276
`https://mcp.atlassian.com/v1/mcp/authv2`, et donne accès aux outils Jira et Confluence. La création de tickets Jira nécessite les autorisations Atlassian permettant l'écriture dans Jira ; l'utilisateur ne voit toujours que les projets et l - Network accessline 278
Les administrateurs d'une organisation Atlassian peuvent décider si un client OAuth est autorisé à se connecter. Avant la connexion, un administrateur peut devoir ajouter l'origine HTTPS réelle de l'application à la liste des domaines de co - Network accessline 284
1. **Nom du serveur** — une courte étiquette pour votre propre référence (par exemple "zapier", "slack-tools"). 2. **URL** : le point de terminaison HTTPS que le fournisseur de serveur MCP vous a fourni (par exemple, `https://mcp.zapier.com - Network accessline 285
2. **URL** : le point de terminaison HTTPS que le fournisseur de serveur MCP vous a fourni (par exemple, `https://mcp.zapier.com/s/abc123/mcp`). Cela se trouve généralement dans le tableau de bord du fournisseur ou dans les documents d'inté - Network accessline 300
[Credential-related source omitted] - Network accessline 394
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 395
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 396
"url": "https://example.com/mcp", - Network accessline 434
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com
packages/core/docs/content/locales/fr-FR/mcp-protocol.mdx
- Access to secretsline 31
[Credential-related source omitted] - Network accessline 108
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 117
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 118
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 119
"url": "https://mail.example.com/mcp", - Network accessline 138
POST https://your-app.example.com/mcp - Access to secretsline 143
[Credential-related source omitted] - Access to secretsline 256
[Credential-related source omitted] - Network accessline 264
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 270
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 271
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 386
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 395
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 396
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 397
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/locales/fr-FR/messaging.mdx
- Network accessline 81
1. Accédez à **[api.slack.com/apps](https://api.slack.com/apps)** et cliquez sur **Créer une nouvelle application** → **À partir de zéro**. Nommez-le (par exemple "Agent") et choisissez votre espace de travail. - Network accessline 97
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 120
https://your-clips.example.com/api/slack/unfurl - Access to secretsline 134
[Credential-related source omitted] - Network accessline 169
Définissez OAuth Redirect URL sur `https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 170
`https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 171
Event Subscriptions Request URL sur `https://your-app.example.com/_agent-native/integrations/slack/webhook` et Interactivity Request URL sur - Network accessline 172
`https://your-app.example.com/_agent-native/integrations/slack/webhook` et - Network accessline 173
Interactivity Request URL sur `https://your-app.example.com/_agent-native/integrations/slack/interactions`. Abonnez-vous à `app_home_opened`, `app_context_changed`, `app_mention`, - Network accessline 174
`https://your-app.example.com/_agent-native/integrations/slack/interactions`. - Access to secretsline 177
[Credential-related source omitted] - Network accessline 273
[Credential-related source omitted] - Access to secretsline 273
[Credential-related source omitted] - Access to secretsline 278
[Credential-related source omitted] - Access to secretsline 375
[Credential-related source omitted] - Network accessline 389
1. Ouvrez le télégramme et envoyez le message **[@BotFather](https://t.me/BotFather)**. - Network accessline 396
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 415
1. Inscrivez-vous sur **[resend.com](https://resend.com)**. L'offre gratuite est suffisante pour commencer. - Network accessline 422
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 429
[Credential-related source omitted] - Network accessline 436
1. Inscrivez-vous sur **[sendgrid.com](https://sendgrid.com)**. - Network accessline 444
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 449
[Credential-related source omitted] - Network accessline 473
1. Accédez à **[Meta Developer Portal](https://developers.facebook.com/)**, cliquez sur **Créer une application** et sélectionnez le type **Business**. - Access to secretsline 480
[Credential-related source omitted] - Network accessline 486
Rappel URL : https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Access to secretsline 697
[Credential-related source omitted] - Access to secretsline 698
[Credential-related source omitted] - Access to secretsline 808
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/multi-app-workspace.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 100
[Credential-related source omitted] - Network accessline 245
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 246
"urls": { "design": "https://design.example.test" } - Network accessline 247
"design": "https://design.example.test" - Access to secretsline 276
[Credential-related source omitted] - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Network accessline 325
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 326
# https://your-agents.com/mail/* → apps/mail - Network accessline 327
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 328
# https://your-agents.com/forms/* → apps/forms
packages/core/docs/content/locales/fr-FR/native-chat-ui.mdx
- Network accessline 453
Consultez le [guide de migration du chat AgentKit](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md) pour connaître les anciennes et nouvelles API ainsi que les étapes de migration.
packages/core/docs/content/locales/fr-FR/netlify.mdx
packages/core/docs/content/locales/fr-FR/node-js.mdx
packages/core/docs/content/locales/fr-FR/notifications.mdx
- Network accessline 192
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 193
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 194
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 211
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 212
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 213
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/locales/fr-FR/observability.mdx
- Network accessline 509
Le dossier [`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel) du dépôt contient une version plus complète de ce plugin, avec des endpoints par signal et une validation de l'exportateur. Il n'est pas publié
packages/core/docs/content/locales/fr-FR/onboarding.mdx
- Access to secretsline 187
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/plan-plugin.mdx
- Network accessline 22
- **Trois compétences** — `/visual-plan` (le point d'entrée canonique), `/visual-recap` et `/visualize-repo`. - **Le connecteur MCP Plan** — enregistré contre l'application hébergée à `https://plan.agent-native.com` (point de terminaison MC - Network accessline 23
- **Le connecteur MCP Plan** — enregistré contre l'application hébergée à `https://plan.agent-native.com` (point de terminaison MCP `https://plan.agent-native.com/mcp`, nom de serveur `plan`). - Network accessline 101
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 130
- **Utilisez un navigateur Chromium.** Safari bloque la page Plan HTTPS hébergée de lire le pont localhost `http://127.0.0.1` (contenu mixte / - Network accessline 131
hébergée de lire le pont localhost `http://127.0.0.1` (contenu mixte / - Network accessline 245
locaux recevoir la configuration MCP (`all` par défaut). - `--no-connect` — enregistre le connecteur sans authentifier ; exécutez plus tard `npx @agent-native/core@latest connect https://plan.agent-native.com --client all`, - Network accessline 246
- `--no-connect` — enregistre le connecteur sans authentifier ; exécutez plus tard `npx @agent-native/core@latest connect https://plan.agent-native.com --client all`, ou choisissez un `--client` plus restreint. - Network accessline 247
plus tard `npx @agent-native/core@latest connect https://plan.agent-native.com --client all`, - Network accessline 314
compétences et les outils MCP se chargent dans la session. Le plugin est livré avec un connecteur URL uniquement (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`) ; `codex mcp login plan` exécute le - Network accessline 315
livré avec un connecteur URL uniquement (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`) ; `codex mcp login plan` exécute le flux OAuth. La route CLI universelle ci-dessus fonctionne aussi pour - Network accessline 316
`https://plan.agent-native.com/mcp`) ; `codex mcp login plan` exécute le - Network accessline 324
**Anciennes installations :** si votre configuration a encore une entrée `agent-native-plans` pointant vers la même URL, exécuter `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 325
`agent-native-plans` pointant vers la même URL, exécuter `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` pour Codex, ou la même commande avec votre `--client` cible, la consolide - Network accessline 326
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`
packages/core/docs/content/locales/fr-FR/pr-visual-recap.mdx
- Access to secretsline 235
[Credential-related source omitted] - Access to secretsline 236
[Credential-related source omitted] - Access to secretsline 237
[Credential-related source omitted] - Network accessline 272
Pour DeepSeek, définissez `VISUAL_RECAP_AGENT=openai-compatible`, `VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`, - Network accessline 273
`VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`, - Access to secretsline 274
[Credential-related source omitted] - Access to secretsline 312
[Credential-related source omitted] - Access to secretsline 344
[Credential-related source omitted] - Network accessline 359
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 377
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 390
[Credential-related source omitted] - Access to secretsline 391
[Credential-related source omitted] - Network accessline 397
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | Variable, variante copie uniquement (l'appelant réutilisable l'ignore). Définissez-la sur `false` pour ne plus générer de recap à chaq - Network accessline 399
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | Secret. Uniquement en cas d'auto-hébergement de l'application Plans sur une autre origine. - Network accessline 449
Pour l'installer, copiez le fichier depuis [BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Network accessline 450
[BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Access to secretsline 453
[Credential-related source omitted] - Access to secretsline 706
[Credential-related source omitted] - Access to secretsline 779
[Credential-related source omitted] - Access to secretsline 780
[Credential-related source omitted] - Access to secretsline 781
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/privacy-and-data-rights.mdx
- Network accessline 59
- [Texte du RGPD](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [Commission européenne : demandes des personnes](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-indi - Network accessline 60
- [Commission européenne : demandes des personnes](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [Attorney General de Californie : CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [Règlements CCPA](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/locales/fr-FR/real-time-collaboration.mdx
- Network accessline 688
`GET /_agent-native/health` rapporte le résultat, et sa résolution effectue l'enregistrement : un seul curl active donc et confirme à la fois. - Network accessline 689
l'enregistrement : un seul curl active donc et confirme à la fois. - Network accessline 692
curl -s https://your-app.example.com/_agent-native/health | jq .realtime
packages/core/docs/content/locales/fr-FR/recurring-jobs.mdx
packages/core/docs/content/locales/fr-FR/render.mdx
packages/core/docs/content/locales/fr-FR/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- N'utilisez jamais `innerHTML`, `eval()` ou `document.write()` - Access to secretsline 395
[Credential-related source omitted] - Access to secretsline 411
[Credential-related source omitted] - Access to secretsline 417
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/server-database.mdx
- Network accessline 8
Toute application native d'agent stocke son état en PostgreSQL. L'UI et l'agent lisent et écrivent tous deux dans les mêmes tables via les mêmes actions, en utilisant le même client [Drizzle ORM](https://orm.drizzle.team). La synchronisatio - Network accessline 81
[PGlite](https://pglite.dev/) exécute une base de données Postgres WASM en cours de processus. Elle vous permet de détecter les problèmes de schéma propres à Postgres avant le déploiement, tout en gardant une configuration aussi simple que
packages/core/docs/content/locales/fr-FR/server-middleware.mdx
- Network accessline 106
Pour d'autres modèles de middleware — logique conditionnelle, opérations asynchrones, interception de réponse — consultez la [documentation middleware de Nitro](https://nitro.build/guide/routing#middleware) et la [documentation H3](https:// - Network accessline 149
- [**Routes**](/docs/server-routes) : les routes personnalisées protégées par ce middleware - [**Security**](/docs/security) : cadrage des données, gardes d'accès, et le modèle d'authentification complet - [**Documentation middleware de Nit - Network accessline 151
- [**Documentation middleware de Nitro**](https://nitro.build/guide/routing#middleware) : l'API et les modèles complets de middleware
packages/core/docs/content/locales/fr-FR/server-overview.mdx
- Network accessline 206
[Nitro](https://nitro.build) est la boîte à outils serveur qui alimente la couche HTTP des applications natives d'agent. La majeure partie de ce que vous configurerez dans `server/` relève de Nitro : routes basées sur des fichiers, plugins - Network accessline 214
L'API de requête/réponse sous-jacente est [H3](https://h3.unjs.io). Lorsque vous voyez `defineEventHandler`, `createError` ou `getQuery` dans des routes personnalisées, il s'agit de H3. La [documentation Nitro](https://nitro.build/guide) et
packages/core/docs/content/locales/fr-FR/server-plugins.mdx
- Network accessline 141
Les plugins peuvent utiliser n'importe lequel des hooks de cycle de vie de Nitro : `request`, `beforeResponse`, `close`, et d'autres. La [documentation des plugins Nitro](https://nitro.build/guide/plugins) couvre en détail l'API complète de - Network accessline 163
- [**Routes**](/docs/server-routes) : les routes de fichiers personnalisées et quand les utiliser - [**Writing Agent Instructions**](/docs/writing-agent-instructions) : comment concevoir le `systemPrompt` dans le plugin agent-chat - [**Docu - Network accessline 165
- [**Documentation des plugins Nitro**](https://nitro.build/guide/plugins) : l'API complète des plugins et les hooks de cycle de vie
packages/core/docs/content/locales/fr-FR/server-routes.mdx
packages/core/docs/content/locales/fr-FR/skills-guide.mdx
- Network accessline 315
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 316
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 317
"url": "https://assets.agent-native.com", - Network accessline 318
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/locales/fr-FR/template-analytics.mdx
packages/core/docs/content/locales/fr-FR/template-assets.mdx
- Network accessline 119
Démo en direct : [assets.agent-native.com](https://assets.agent-native.com). - Network accessline 468
<EmbeddedApp url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 469
url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 505
La compétence d'application Actifs porte l'ID d'application `assets` et héberge MCP URL `https://assets.agent-native.com/mcp`. - Network accessline 506
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/locales/fr-FR/template-brain.mdx
- Network accessline 159
Démo en direct : [brain.agent-native.com](https://brain.agent-native.com).
packages/core/docs/content/locales/fr-FR/template-calendar-developers.mdx
- Network accessline 26
Cela nécessite [Node.js](https://nodejs.org) 22.22 ou une version plus récente, et - Network accessline 27
[pnpm](https://pnpm.io) installé au préalable. Consultez - Network accessline 38
L'application s'ouvre à `http://localhost:8082`. Si elle ne s'ouvre pas automatiquement, - Network accessline 49
1. **Créer un projet Google Cloud**, si vous n'en avez pas déjà un à utiliser : dans la [Console Google Cloud](https://console.cloud.google.com/), cliquez sur le - Network accessline 50
dans la [Console Google Cloud](https://console.cloud.google.com/), cliquez sur le - Network accessline 59
4. **Créer des identifiants** : sous **APIs & Services → Identifiants**, créez un ID client OAuth de type « Application Web », et ajoutez `http://localhost:8082/_agent-native/google/callback` comme URI de - Network accessline 60
un ID client OAuth de type « Application Web », et ajoutez `http://localhost:8082/_agent-native/google/callback` comme URI de redirection autorisée (8082 est le port de développement par défaut de Calendrier). - Network accessline 61
`http://localhost:8082/_agent-native/google/callback` comme URI de - Access to secretsline 78
[Credential-related source omitted] - Network accessline 81
Obtenez un ID client et un secret depuis une [application OAuth Zoom](https://marketplace.zoom.us/)
packages/core/docs/content/locales/fr-FR/template-calendar-features.mdx
packages/core/docs/content/locales/fr-FR/template-calendar-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/template-calendar.mdx
packages/core/docs/content/locales/fr-FR/template-chat-developers.mdx
- Network accessline 31
Cela nécessite [Node.js](https://nodejs.org) 22.22 ou une version plus récente. Si `pnpm` n'est pas dans
packages/core/docs/content/locales/fr-FR/template-chat-integrations.mdx
- Access to secretsline 22
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/template-chat.mdx
packages/core/docs/content/locales/fr-FR/template-clips-developers.mdx
- Network accessline 22
Cela nécessite [Node.js](https://nodejs.org) 22.22 ou une version ultérieure et [pnpm](https://pnpm.io) installés au préalable. Consultez [Bien démarrer](/docs/getting-started) si vous n'avez encore installé aucun des deux, ou si vous n'ave - Network accessline 310
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 311
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 312
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/template-clips-embed.mdx
- Network accessline 40
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 53
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 54
[Credential-related source omitted] - Network accessline 56
const response = await fetch( - Network accessline 145
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 152
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/locales/fr-FR/template-clips-features.mdx
- Network accessline 117
L'application de bureau est conçue pour la capture quotidienne la plus fluide : enregistrement d'écran et de webcam avec une bulle de caméra déplaçable, transcription en direct pendant que vous enregistrez, tout cela sans garder un onglet d - Network accessline 212
[Credential-related source omitted] - Network accessline 213
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/template-clips-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/template-clips.mdx
packages/core/docs/content/locales/fr-FR/template-content.mdx
- Network accessline 12
### [Essayez-le maintenant →](https://content.agent-native.com)
packages/core/docs/content/locales/fr-FR/template-design-developers.mdx
packages/core/docs/content/locales/fr-FR/template-design-integrations.mdx
- Access to secretsline 26
[Credential-related source omitted] - Network accessline 53
https://design.agent-native.com/mcp - Network accessline 58
exécute `/mcp` pour s'authentifier. Codex et les autres clients CLI ou stdio exécutent plutôt `npx @agent-native/core@latest connect https://design.agent-native.com`, qui ouvre un onglet de navigateur pour - Network accessline 59
exécutent plutôt `npx @agent-native/core@latest connect https://design.agent-native.com`, qui ouvre un onglet de navigateur pour autoriser et génère un jeton révocable par utilisateur, sans aucun secret - Network accessline 60
https://design.agent-native.com`, qui ouvre un onglet de navigateur pour - Network accessline 76
erreur de session ou d'authentification, n'insistez pas en le relançant. Exécutez `/mcp` dans Claude Code et reconnectez-vous, ou exécutez `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` depuis - Network accessline 77
Exécutez `/mcp` dans Claude Code et reconnectez-vous, ou exécutez `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` depuis n'importe quel terminal. Les deux réauthentifient sans rien réinstaller. - Network accessline 78
@agent-native/core@latest reconnect https://design.agent-native.com` depuis
packages/core/docs/content/locales/fr-FR/template-design.mdx
packages/core/docs/content/locales/fr-FR/template-dispatch.mdx
packages/core/docs/content/locales/fr-FR/template-forms-agent.mdx
- Network accessline 40
- **« Publie ceci et donne-moi le lien. »** Le formulaire passe en ligne et vous obtenez son URL publique tout de suite. - **« Crée un formulaire de retour client anonyme et donne-moi le lien public. »** Le formulaire est publié anonyme dès - Network accessline 42
- **« N'autorise ce formulaire à être soumis que depuis https://example.com. »** Le formulaire commence à rejeter les soumissions venant de partout ailleurs. Ceci n'est disponible qu'en le demandant à l'agent aujourd'hui. Il n'y a pas d'int
packages/core/docs/content/locales/fr-FR/template-forms-features.mdx
- Network accessline 14
[Credential-related source omitted] - Network accessline 15
[Credential-related source omitted] - Network accessline 94
Les soumissions publiques peuvent nécessiter une vérification via [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/) avant de passer. Activez ceci en définissant une clé de site et une clé secrète dans Paramètres. Deux p - Network accessline 110
Il n'y a pas d'interrupteur pour cela dans le générateur aujourd'hui. Demandez à l'agent de le configurer, par exemple « n'autorise ce formulaire à être soumis que depuis https://example.com ». Donnez à l'agent l'origine exacte (le schéma e - Network accessline 116
[Credential-related source omitted] - Network accessline 117
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/template-forms.mdx
packages/core/docs/content/locales/fr-FR/template-mail.mdx
- Network accessline 97
Démo en direct : [mail.agent-native.com](https://mail.agent-native.com). - Network accessline 168
1. Ouvrez [Google Cloud Console](https://console.cloud.google.com/) et créez un projet. - Network accessline 169
2. Activez **Gmail API** sous APIs & Services → Bibliothèque. 3. Créez les identifiants OAuth 2.0 (type : application Web). Ajoutez `http://localhost:8085/_agent-native/google/callback` comme redirection autorisée URI. 4. Copiez l'ID client - Network accessline 170
3. Créez les identifiants OAuth 2.0 (type : application Web). Ajoutez `http://localhost:8085/_agent-native/google/callback` comme redirection autorisée URI.
packages/core/docs/content/locales/fr-FR/template-plan.mdx
- Network accessline 145
`/visualize-repo` — ainsi que le connecteur MCP `plan` (un serveur d'outils distant avec lequel votre agent dialogue) sur `https://plan.agent-native.com/mcp`. Pour la référence complète des - Network accessline 146
distant avec lequel votre agent dialogue) sur `https://plan.agent-native.com/mcp`. Pour la référence complète des options, les routes d'installation par plugin/marketplace pour Claude Code - Network accessline 147
`https://plan.agent-native.com/mcp`. Pour la référence complète des - Network accessline 246
Si un outil Plans renvoie un jour `needs auth`, `Unauthorized`, ou `Session terminated`, ne continuez pas à réessayer. Réauthentifiez-vous avec `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, - Network accessline 247
terminated`, ne continuez pas à réessayer. Réauthentifiez-vous avec `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, ou relancez `/mcp` → **Authenticate** pour un hôte compatible OAuth. - Network accessline 248
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`,
packages/core/docs/content/locales/fr-FR/template-slides-features.mdx
packages/core/docs/content/locales/fr-FR/template-slides-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/template-slides.mdx
packages/core/docs/content/locales/fr-FR/toolkit-capability-packages.mdx
packages/core/docs/content/locales/fr-FR/toolkit-settings.mdx
- Access to secretsline 121
[Credential-related source omitted] - Network accessline 191
Cette implémentation nécessite un rappel OAuth HTTP sur `127.0.0.1`, l’adresse de bouclage sur l’ordinateur de l’utilisateur ; `localhost` ne peut pas la remplacer. Ce flux en libre-service ne fonctionne pas sur les sites hébergés `*.agent- - Network accessline 201
Utilisez uniquement la Responses API publique pour l’inférence. Les points de terminaison privés du backend de ChatGPT et les données de ses conversations ne sont pas couverts par cette autorisation. OpenAI indique que le partage de jetons
packages/core/docs/content/locales/fr-FR/tracking.mdx
- Network accessline 83
| Fournisseur | Vars d'environnement | | ----------------- | ---------------------------------------------------------------------------------------------- - Network accessline 84
| ----------------- | --------------------------------------------------------------------------------------------------- | | PostHog | `POSTHOG_API_KEY` (obligatoire), `POSTHOG_HOST` (facultatif, par défaut `https://us.i.posthog. - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Network accessline 205
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 206
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 207
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 225
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 226
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 283
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 284
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 285
fetch("https://analytics.example.com/events", { - Network accessline 343
- **HTTP par lots** : les fournisseurs intégrés mettent les événements en file d'attente et les vident toutes les 10 secondes ou lorsque 50 événements s'accumulent, selon la première éventualité. Cela minimise les requêtes sortantes sans pe - Network accessline 344
- **Aucune dépendance SDK** — tous les fournisseurs intégrés utilisent `fetch()` brut. Pas de PostHog SDK, pas de Mixpanel SDK, pas d'Amplitude SDK. Maintient la légèreté du framework.
packages/core/docs/content/locales/fr-FR/turn-into-app.mdx
- Network accessline 12
- Un agent d'IA capable d'exécuter des commandes de terminal, comme Claude Code, Codex ou Cursor. Claude ou ChatGPT sur le web peut à la place confier le travail à Builder. Voir [Où elle s'exécute](#hosts). - [Node.js](https://nodejs.org). - Network accessline 13
- [Node.js](https://nodejs.org). L'installateur de nodejs.org inclut `npx`, qui exécute les commandes d'installation ci-dessous. - Network accessline 18
- **Sur votre machine.** Node.js 22.22 ou version ultérieure (vérifiez avec `node --version`) et [pnpm](https://pnpm.io) dans votre `PATH`. Si `pnpm --version` échoue, exécutez `corepack enable`, fourni avec Node.js. - Network accessline 59
L'hébergement par Builder, depuis l'un ou l'autre type d'hôte, exige le connecteur MCP Agent-Native Dispatch (`https://dispatch.agent-native.com/mcp`, OAuth) et appelle son action `start-workspace-app-creation`. Rien ne s'exécute de votre c - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/locales/fr-FR/updating-ui-in-production.mdx
- Network accessline 18
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment) est un workflow géré séparé et capable de coder où un agent peut modifier l'interface de votre application en product
packages/core/docs/content/locales/fr-FR/upgrading-core-ui.mdx
- Access to secretsline 77
[Credential-related source omitted] - Access to secretsline 80
[Credential-related source omitted] - Network accessline 196
URL canonique de la source : [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx). - Network accessline 197
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx).
packages/core/docs/content/locales/fr-FR/vercel.mdx
- Access to secretsline 103
[Credential-related source omitted] - Network accessline 122
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 150
[Credential-related source omitted] - Network accessline 218
d’attente et le nettoyage des exécutions interrompues passent tous par un balayage exécuté une fois par minute. Le build ajoute une [tâche cron Vercel](https://vercel.com/docs/cron-jobs) qui appelle le balayage à - Network accessline 219
balayage exécuté une fois par minute. Le build ajoute une [tâche cron Vercel](https://vercel.com/docs/cron-jobs) qui appelle le balayage à `/_agent-native/jobs/_process-sweep`. Un build de workspace ajoute une tâche - Network accessline 220
[tâche cron Vercel](https://vercel.com/docs/cron-jobs) qui appelle le balayage à
packages/core/docs/content/locales/fr-FR/voice-input.mdx
packages/core/docs/content/locales/fr-FR/what-is-agent-native.mdx
packages/core/docs/content/locales/fr-FR/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Network accessline 271
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 272
[Credential-related source omitted] - Network accessline 273
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 274
[Credential-related source omitted] - Access to secretsline 275
[Credential-related source omitted] - Access to secretsline 276
[Credential-related source omitted] - Network accessline 288
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 289
[Credential-related source omitted] - Network accessline 290
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 291
[Credential-related source omitted] - Access to secretsline 292
[Credential-related source omitted] - Access to secretsline 293
[Credential-related source omitted] - Network accessline 344
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 346
APP_URL=https://workspace.example.com
packages/core/docs/content/locales/fr-FR/workspace-management.mdx
- Network accessline 90
Tous ceux qui ont besoin d'apporter des modifications ne sont pas à l'aise avec git. [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management) prend en charge un mod - Network accessline 140
Conseils clés : utilisez les équipes GitHub (`@org/team`), et non les individus. Les modifications du framework et de la répartition devraient toujours nécessiter un examen de la plate-forme. Voir [GitHub CODEOWNERS docs](https://docs.githu - Network accessline 160
Ajoutez ensuite l'action [actions/labeler](https://github.com/actions/labeler) — consultez le README de ce dépôt pour le flux de travail complet YAML. Les étiquettes s'appliquent automatiquement lorsque les PR sont ouvertes ou mises à jour. - Access to secretsline 233
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/a2a-protocol.mdx
- Access to secretsline 93
[Credential-related source omitted] - Network accessline 117
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 118
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 119
"url": "https://analytics.example.com", - Network accessline 314
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 350
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 351
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 352
"https://analytics.example.com", - Access to secretsline 354
[Credential-related source omitted] - Network accessline 372
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", - Network accessline 373
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", "cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 374
"url": "https://example.services.ai.azure.com/a2a", - Network accessline 375
"cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 412
[Azure AI Foundry hosted agents](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 418
Foundry [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint) - Network accessline 428
[Gemini Enterprise external A2A agents को register कर सकता है](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent) - Network accessline 431
invocation के लिए `POST {url}/v1/message:send` और streaming के लिए `POST {url}/v1/message:stream`। Requests Google OAuth या ADC bearer tokens और Google Cloud IAM का उपयोग करती हैं। Endpoint और permissions के लिए [registry A2A endpoint से ag - Network accessline 433
Google Cloud IAM का उपयोग करती हैं। Endpoint और permissions के लिए [registry A2A endpoint से agent को call करें](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a)। - Network accessline 449
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 450
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 451
"url": "https://api.anthropic.com", - Access to secretsline 456
[Credential-related source omitted] - Access to secretsline 469
[Credential-related source omitted] - Network accessline 515
# किसी डिप्लॉय किए गए ऐप को सीधे URL से कॉल करें, bearer token env से पढ़ते हुए agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 516
[Credential-related source omitted] - Access to secretsline 516
[Credential-related source omitted] - Access to secretsline 517
[Credential-related source omitted] - Access to secretsline 589
[Credential-related source omitted] - Access to secretsline 593
[Credential-related source omitted] - Network accessline 651
- एक साझा `A2A_SECRET` को निर्माण के समय प्रत्येक ऐप के वातावरण में माउंट किया जाता है। - क्रॉस-ऐप कॉल समान-मूल हैं - `https://workspace.example.com/apps/analytics` कॉल `https://workspace.example.com/apps/mail` - इसलिए कोई DNS, CORS, या प्र - Network accessline 652
- क्रॉस-ऐप कॉल समान-मूल हैं - `https://workspace.example.com/apps/analytics` कॉल `https://workspace.example.com/apps/mail` - इसलिए कोई DNS, CORS, या प्रति-जोड़ी JWT सेटअप नहीं है। - Network accessline 660
**कभी भी प्रतिक्रिया को समाप्त करने वाले `Promise` पर भरोसा न करें। ** सर्वर रहित फ़ंक्शंस (नेटलिफाई, वर्सेल, AWS लैम्ब्डा, क्लाउड रन) प्रतिक्रिया निकाय के फ्लश होते ही रुक जाते हैं - कभी-कभी एक अप्रत्याशित `fetch(...)` के TCP हैंडशेक के पू - Network accessline 698
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 699
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 700
"https://analytics.example.com", - Access to secretsline 702
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/actions-defining.mdx
- Network accessline 226
`schema` कोई भी [Standard Schema](https://standardschema.dev)-संगत लाइब्रेरी स्वीकार करता है:
packages/core/docs/content/locales/hi-IN/add-ui-to-your-agent.mdx
- Network accessline 27
A2A v1.0 Agent Card URL डालें और Entra credential चुनें। Service-to-service के लिए `https://ai.azure.com/.default` scope और **Foundry Agent Consumer** role दें। - Network accessline 45
[Global A2A Registry](https://www.a2a-registry.org) browse करें। Agent जोड़ने और secret इस्तेमाल करने से पहले card, host और auth verify करें।
packages/core/docs/content/locales/hi-IN/agent-mentions.mdx
packages/core/docs/content/locales/hi-IN/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/agent-web-surfaces.mdx
- Network accessline 171
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 172
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 173
siteUrl: "https://example.com", - Network accessline 198
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/locales/hi-IN/audit-log.mdx
- Network accessline 110
**`export-audit-ocsf`** संगठन के ट्रेल को [OCSF](https://schema.ocsf.io) 1.9.0 API Activity इवेंट (`class_uid` 6003) के रूप में सबसे पुराने से शुरू करके लौटाता है, ताकि SIEM उसे नियमित रूप से खींच सके। केवल ओनर और एडमिन के लिए; यह संगठन के
packages/core/docs/content/locales/hi-IN/authentication.mdx
- Network accessline 8
एजेंट-नेटिव ऐप्स अकाउंट-फर्स्ट डिज़ाइन के साथ प्रमाणीकरण के लिए [Better Auth](https://better-auth.com) का उपयोग करते हैं। उपयोगकर्ता पहली विज़िट पर एक खाता बनाते हैं और पहले दिन से वास्तविक पहचान प्राप्त करते हैं। - Access to secretsline 195
[Credential-related source omitted] - Access to secretsline 200
[Credential-related source omitted] - Access to secretsline 204
[Credential-related source omitted] - Access to secretsline 209
[Credential-related source omitted] - Access to secretsline 211
[Credential-related source omitted] - Access to secretsline 233
[Credential-related source omitted] - Access to secretsline 237
[Credential-related source omitted] - Network accessline 250
https://mail.agent-native.com/mcp - Access to secretsline 475
[Credential-related source omitted] - Access to secretsline 477
[Credential-related source omitted] - Access to secretsline 479
[Credential-related source omitted] - Access to secretsline 480
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/automations.mdx
packages/core/docs/content/locales/hi-IN/aws-amplify.mdx
- Access to secretsline 115
[Credential-related source omitted] - Network accessline 132
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 194
[Credential-related source omitted] - Network accessline 208
[Amplify console](https://console.aws.amazon.com/amplify/) में ऐप बनाएं: - Network accessline 243
पहले build में कुछ मिनट लगते हैं, और आप Amplify console में उसे देख सकते हैं। पूरा होने पर ऐप का address खोलें, जो `https://main.d1m7bkiki6tdw1.amplifyapp.com` जैसा दिखता है, और पहला account - Network accessline 244
पूरा होने पर ऐप का address खोलें, जो `https://main.d1m7bkiki6tdw1.amplifyapp.com` जैसा दिखता है, और पहला account बनाएं। - Network accessline 245
`https://main.d1m7bkiki6tdw1.amplifyapp.com` जैसा दिखता है, और पहला account - Access to secretsline 300
[Credential-related source omitted] - Access to secretsline 304
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/aws-lambda.mdx
- Network accessline 38
आपको एक AWS account भी चाहिए, और अपने computer पर installed और signed-in [AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html) - Network accessline 39
[AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html) - Access to secretsline 116
[Credential-related source omitted] - Access to secretsline 151
[Credential-related source omitted] - Access to secretsline 169
[Credential-related source omitted] - Access to secretsline 217
[Credential-related source omitted] - Access to secretsline 225
[Credential-related source omitted] - Network accessline 250
पहला command `FunctionUrl` print करता है, जो `https://abc123.lambda-url.us-east-1.on.aws/` जैसा दिखता है। इसे चरण 7 के लिए copy करें। दोनों `add-permission` commands requests को उस address के ज़रिए function तक - Network accessline 251
`https://abc123.lambda-url.us-east-1.on.aws/` जैसा दिखता है। इसे चरण 7 के लिए copy करें। दोनों `add-permission` commands requests को उस address के ज़रिए function तक पहुंचने देते हैं। - Network accessline 270
[CloudFront console](https://console.aws.amazon.com/cloudfront/v4/home) में - Network accessline 285
Distribution के **Origins** tab पर, एक origin बनाएं। **Origin domain** के लिए, चरण 5 के function URL का host name paste करें, `https://` या आखिरी slash के बिना। - Network accessline 286
चरण 5 के function URL का host name paste करें, `https://` या आखिरी slash के बिना। - Access to secretsline 311
[Credential-related source omitted] - Network accessline 321
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 323
[Credential-related source omitted] - Access to secretsline 325
[Credential-related source omitted] - Access to secretsline 342
[Credential-related source omitted] - Dangerous shell operationline 344
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/locales/hi-IN/azure-static-web-apps.mdx
packages/core/docs/content/locales/hi-IN/blueprint-installer.mdx
- Network accessline 68
agent-native add <kind> <name> # print a curated blueprint agent-native add <kind> <https://docs…> # URL से अनुसंधान-और-एकीकृत करें - Network accessline 69
agent-native add <kind> <https://docs…> # URL से अनुसंधान-और-एकीकृत करें - Network accessline 95
agent-native add provider https://docs.example.com/api | claude
packages/core/docs/content/locales/hi-IN/cli-adapters.mdx
- Access to secretsline 99
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/locales/hi-IN/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">घटक</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 126
optional: true, description: "[useQuery options](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) से सब कुछ, `queryKey` और `queryFn` को छोड़कर, जिन्हें हुक स्वयं सेट करता है।", - Network accessline 127
description: "[useQuery options](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) से सब कुछ, `queryKey` और `queryFn` को छोड़कर, जिन्हें हुक स्वयं सेट करता है।", }, - Network accessline 128
"[useQuery options](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) से सब कुछ, `queryKey` और `queryFn` को छोड़कर, जिन्हें हुक स्वयं सेट करता है।", - Network accessline 138
यह हुक केवल-पढ़ने योग्य actions के लिए उपयोग किए जाने का इरादा है। यह [useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) फ़ंक्शन के इर्द-गिर्द एक पतला रैपर है। - Network accessline 174
optional: true, description: "[useMutation options](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) से सब कुछ, `mutationFn` को छोड़कर, जिसे हुक स्वयं सेट करता है।", - Network accessline 175
description: "[useMutation options](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) से सब कुछ, `mutationFn` को छोड़कर, जिसे हुक स्वयं सेट करता है।", fields: [ - Network accessline 176
"[useMutation options](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) से सब कुछ, `mutationFn` को छोड़कर, जिसे हुक स्वयं सेट करता है।", - Network accessline 202
यह हुक उन actions के लिए उपयोग किए जाने का इरादा है जो डेटा बदलती हैं, जैसे बनाना, अपडेट करना, या कुछ भेजना। यह [useMutation()](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) फ़ंक्शन के इर्द-गिर्द एक पतला रैपर
packages/core/docs/content/locales/hi-IN/cloneable-saas.mdx
- Network accessline 48
https://github.com/owner/repo - Network accessline 65
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 66
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 74
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 75
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 145
सूची में आना चाहते हैं? रिपॉज़िटरी URL, एक वाक्य का विवरण, मेंटेनर और लाइसेंस जानकारी, टेस्ट किया हुआ CLI कमांड और वैकल्पिक hosted demo URL के साथ [कम्युनिटी टेम्पलेट सबमिट करें](https://github.com/BuilderIO/agent-native/issues/new?template - Network accessline 147
[कम्युनिटी टेम्पलेट सबमिट करें](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml)। - Access to secretsline 299
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/cloudflare.mdx
- Access to secretsline 100
[Credential-related source omitted] - Network accessline 117
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 166
[Credential-related source omitted] - Network accessline 210
हर मिनट एक बार चलने वाले sweep पर होते हैं। Build generated `wrangler.json` में एक `* * * * *` [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) - Network accessline 211
एक `* * * * *` [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) जोड़ता है और आपके पहले से declare किए गए Cron Triggers को बनाए रखता है। जब यह - Network accessline 212
[Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/)
packages/core/docs/content/locales/hi-IN/code-agents-ui.mdx
- Network accessline 202
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 203
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 355
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 356
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Possible obfuscationline 433
<Diagram id="doc-block-kmpho0" title={"रिमोट Dispatch केवल आउटबाउंड है"} summary={"मोबाइल कभी भी डेस्कटॉप से <U+200B><U+200B>सीधे बात नहीं करता. Desktop लंबे-चुनाव Dispatch, कमांड का दावा करता है, स्थानीय रन स्टोर चलाता है, और परिणामों को वापस दिखाता है।
packages/core/docs/content/locales/hi-IN/create-with-an-agent.mdx
- Network accessline 14
- ऐसा AI agent जो terminal commands चला सके, जैसे Claude Code, Codex या Cursor। जो agents सिर्फ chat करते हैं, जैसे ChatGPT या web पर Claude, वे भी मदद कर सकते हैं। [Commands न चला सकने वाले agents](#chat-only-agents) देखें। - [Node.js](htt - Network accessline 15
- [Node.js](https://nodejs.org) 22.22 या बाद का संस्करण। - Network accessline 40
5. `pnpm dev` से app चालू करो और उसे background में चलता रखो। 6. `pnpm dev` जो local URL प्रिंट करे उसे पढ़ो। जांचो कि वह जवाब देता है, जैसे `curl -I <url>` से। browser मत खोलो और - Network accessline 41
6. `pnpm dev` जो local URL प्रिंट करे उसे पढ़ो। जांचो कि वह जवाब देता है, जैसे `curl -I <url>` से। browser मत खोलो और headless browser tools इस्तेमाल मत करो। app मैं खुद खोलूंगा। - Network accessline 42
जवाब देता है, जैसे `curl -I <url>` से। browser मत खोलो और - Access to secretsline 51
[Credential-related source omitted] - Network accessline 60
- एक नया app folder, जिसका नाम आपके बदले बिना `my-app` है, और जिसमें `actions/`, `app/` और एक `AGENTS.md` file हो। - `pnpm dev` बिना errors के चल रहा हो। - एक local URL, आमतौर पर `http://localhost:8080`, जिसके जवाब देने की agent ने जांच कर - Network accessline 62
- एक local URL, आमतौर पर `http://localhost:8080`, जिसके जवाब देने की agent ने जांच कर ली हो।
packages/core/docs/content/locales/hi-IN/creating-templates.mdx
- Access to secretsline 486
[Credential-related source omitted] - Network accessline 531
रिपॉज़िटरी तैयार होने पर [कम्युनिटी टेम्पलेट सबमिट करें](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml)। कम्युनिटी सूची पहले-पक्ष कैटलॉग से अलग रहती है और उसके लेखक उसका रखरखाव करते हैं।
packages/core/docs/content/locales/hi-IN/cross-app-sso.mdx
- Access to secretsline 61
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/decisions.mdx
- Network accessline 18
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 19
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 20
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/locales/hi-IN/deno-deploy.mdx
- Network accessline 25
आपको एक Deno Deploy account और अपने computer पर installed [Deno](https://docs.deno.com/runtime/getting_started/installation/) - Access to secretsline 99
[Credential-related source omitted] - Network accessline 116
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 168
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/deployment-environment-variables.mdx
- Network accessline 31
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | उन deployments पर on-demand schema DDL skip करें जिन्हें automatic serverless skip cover नहीं करता। ज़रूरी है कि आपका schema पहले से migrate हो चुका हो। देखें [On-demand schema DDL](/docs/depl - Network accessline 33
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | केवल corporate callback के लिए, exact trusted preview origins की comma-separated allowlist (उदाहरण के लिए, `https://0123456789abcdef01234567--content.netlify.app`)। इसे केवल corporate callback - Network accessline 47
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 48
| `BETTER_AUTH_SECRET` | 32+ char की random string। Session cookies sign करती है और `OAUTH_STATE_SECRET` तथा app-local encrypted values की fallback है। यह existing shared vault ciphertext के लिए एक legacy read candida - Network accessline 49
| `BETTER_AUTH_URL` | इस ऐप के लिए वैकल्पिक public origin override (जैसे `https://mail.example.com`)। Framework unset होने पर इसे request, known-template, और hosting-platform context से infer करता है। - Access to secretsline 50
[Credential-related source omitted] - Access to secretsline 51
[Credential-related source omitted] - Access to secretsline 59
[Credential-related source omitted] - Access to secretsline 69
[Credential-related source omitted] - Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/deployment.mdx
packages/core/docs/content/locales/hi-IN/dispatch.mdx
- Network accessline 116
डिस्पैच बाहरी एजेंटों के लिए एकल MCP कनेक्टर हो सकता है: Claude, ChatGPT, Codex, या कर्सर में एक बार `https://dispatch.agent-native.com/mcp` जोड़ें, और प्रति ऐप एक कनेक्टर के बजाय एक प्राधिकरण प्रत्येक स्वीकृत कार्यक्षेत्र ऐप तक पहुंचता है। - Network accessline 125
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p
packages/core/docs/content/locales/hi-IN/docker.mdx
- Dangerous shell operationline 70
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 72
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 113
[Credential-related source omitted] - Access to secretsline 169
[Credential-related source omitted] - Network accessline 186
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/locales/hi-IN/embedding-sdk.mdx
- Network accessline 222
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 223
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 224
url="https://assets.agent-native.com/picker" - Network accessline 277
getMcpUrl("https://assets.agent-native.com"); - Network accessline 278
getA2AUrl("https://assets.agent-native.com"); - Network accessline 280
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 281
"https://assets.agent-native.com", - Network accessline 302
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 303
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 304
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 386
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 388
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 390
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 391
hostOrigin: "https://app.example.com", - Network accessline 394
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 395
hostOrigin: "https://app.example.com", - Network accessline 399
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 400
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 401
{ hostOrigin: "https://app.example.com" }, - Network accessline 405
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 406
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 407
{ hostOrigin: "https://app.example.com" }, - Network accessline 410
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 411
hostOrigin: "https://app.example.com", - Network accessline 464
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 465
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 466
hostOrigin: "https://app.example.com", - Possible obfuscationline 499
क्लाइंट actions एक ब्राउज़र टैब के लिए एक लाइव ब्रिज है। होस्ट उन्हें `source: "client"` और `availability: "browser-session"` के साथ विज्ञापित करता है, और साइडकार को उस अभिव्यक्ति को अस्थायी मानना <U+200B><U+200B>चाहिए। मार्ग या चयन बदलने पर actions को प
packages/core/docs/content/locales/hi-IN/extensions.mdx
- Network accessline 239
| `dbQuery(sql, args)` | SQL से पढ़ें (उपयोगकर्ता के लिए ऑटो-स्कोप) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` | - Network accessline 240
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` | - Network accessline 241
| `extensionFetch(url, options)` | रहस्यों के साथ एक सुरक्षित प्रॉक्सी के माध्यम से बाहरी APIs को हिट करें | `extensionFetch('https://api.github.com/user')` | - Network accessline 269
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 270
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/locales/hi-IN/external-agents.mdx
- Network accessline 111
https://dispatch.agent-native.com/mcp - Network accessline 122
https://mail.agent-native.com/mcp - Network accessline 126
प्रत्येक होस्ट किए गए ऐप में एक सहायक पृष्ठ भी होता है `https://<app>/mcp/connect` प्रतिलिपि योग्य URL और - Network accessline 127
`https://<app>/mcp/connect` प्रतिलिपि योग्य URL और - Network accessline 228
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 272
कच्चा `skills` CLI केवल `SKILL.md` फ़ाइलें स्थापित करता है; स्थानीय MCP ग्राहक अभी भी `npx @agent-native/core@latest connect https://assets.agent-native.com` जैसे कनेक्टर की आवश्यकता है। - Network accessline 273
`npx @agent-native/core@latest connect https://assets.agent-native.com` जैसे कनेक्टर की आवश्यकता है। - Network accessline 279
डिफ़ॉल्ट क्लाइंट चयन सभी स्थानीय क्लाइंट समर्थित है; सेटअप को संकीर्ण करने के लिए `--client codex`, `--client claude-code`, या कोई अन्य विशिष्ट लक्ष्य जोड़ें। इनलाइन होस्ट (ChatGPT, वेब, डेस्कटॉप और मोबाइल पर Claude चैट तथा Claude Desktop क - Network accessline 285
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 300
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 303
`reconnect` किसी भी MCP कॉन्फ़िगरेशन प्रविष्टि को ढूंढता है जिसका URL दिए गए होस्ट और चयनित क्लाइंट के लिए `/mcp` में समाप्त होता है (कनेक्टर नाम की परवाह किए बिना URL द्वारा मिलान), फिर आपके स्थापित skills को छूने या पूर्ण इंस्टॉल प्रवाह क - Network accessline 311
MCP क्लाइंट के लिए जो सीधे रिमोट OAuth URL नहीं जोड़ सकते हैं, अपने ब्राउज़र में ऐप खोलें और इसके **कनेक्ट** अफोर्डेंस (`https://<app>/mcp/connect` पर उपलब्ध) का उपयोग करें। लॉग इन करते समय, **कनेक्ट/अधिकृत** पर क्लिक करें। पेज आपको य - Network accessline 317
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 318
[Credential-related source omitted] - Network accessline 319
"url": "https://mail.agent-native.com/mcp", - Network accessline 335
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 336
https://dispatch.agent-native.com/mcp - Network accessline 339
यह वही URL-केवल प्रविष्टि है जिसे `npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code` आपके लिए लिखता है। फिर Claude कोड में `/mcp` चलाएं और **प्रमाणित करें** चुनें। क्लाइंट MCP सर्वर के `401 WWW-Au - Network accessline 342
https://dispatch.agent-native.com/mcp - Network accessline 483
npx @agent-native/core@latest कनेक्ट https://plan.agent-native.com --क्लाइंट कोडेक्स --फुल-कैटलॉग - Network accessline 534
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 535
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Access to secretsline 766
[Credential-related source omitted] - Access to secretsline 782
[Credential-related source omitted] - Access to secretsline 788
[Credential-related source omitted] - Network accessline 794
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 795
[Credential-related source omitted] - Network accessline 796
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 797
[Credential-related source omitted] - Network accessline 819
`connect dev` स्थानीय देव-आलसी गेटवे पर समान स्थिर MCP सर्वर नाम (`agent-native-mail`, `agent-native-calendar`, आदि) को फिर से लिखता है, इसलिए टूल नाम नहीं बदलते हैं। यह डेव प्रविष्टियाँ लिखने से पहले `~/.agent-native/connect-profiles.json` - Network accessline 846
- `npx @agent-native/core@latest connect https://dispatch.agent-native.com` के साथ प्रेषण के लिए अपने स्वयं के एजेंट को कनेक्ट करें; डायरेक्ट ऐप URL का उपयोग केवल तभी करें जब आपको एक अलग ऐप चाहिए। - Access to secretsline 866
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/file-uploads.mdx
- Network accessline 131
किसी अलग तरह के बैकएंड के लिए ही एक कस्टम प्रदाता पंजीकृत करें। यह उदाहरण `fetch()` का उपयोग करके एक सामान्य HTTP API के माध्यम से अपलोड करता है: - Network accessline 148
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 149
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 150
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/locales/hi-IN/getting-started.mdx
- Network accessline 22
- [Node.js](https://nodejs.org) 22.22 या बाद का संस्करण - Network accessline 23
- [pnpm](https://pnpm.io) इंस्टॉल हो और आपके `PATH` में उपलब्ध हो - Network accessline 150
नया route खोजने के लिए development server restart करें, फिर `http://localhost:<port>/hello` खोलें। - Network accessline 151
`http://localhost:<port>/hello` खोलें।
packages/core/docs/content/locales/hi-IN/harness-agents.mdx
- Network accessline 111
Agent-Native एक [ACP](https://agentclientprotocol.com) (एजेंट क्लाइंट) के रूप में कार्य कर सकता है
packages/core/docs/content/locales/hi-IN/key-concepts.mdx
- Network accessline 325
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 326
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 327
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/locales/hi-IN/koyeb.mdx
packages/core/docs/content/locales/hi-IN/mcp-apps.mdx
- Network accessline 23
OpenAI Plugin Extensions, मॉडल द्वारा बुलाए जाने वाले MCP टूल के साथ ऐप खोलने के बिंदु जोड़ते हैं। Inline MCP Apps सक्षम होने पर साझा `open_app` टूल वैश्विक साइडबार और बातचीत के साइड पैनल, दोनों के लिए entrypoint देता है। ChatGPT इसे `{}` क - Network accessline 261
MCP Apps को इनलाइन दिखाने वाले होस्ट में वेब, डेस्कटॉप और मोबाइल पर Claude चैट तथा Claude Desktop का [Code](https://github.com/anthropics/claude-code/issues/103809) टैब शामिल हैं; ChatGPT; VS Code GitHub Copilot (1.109+); Microsoft 365 Copi - Network accessline 263
Claude Code टर्मिनल, Cowork, claude.ai वेब पर Code और Codex CLI में डीप-लिंक फ़ॉलबैक का उपयोग करें। [Cowork अभी MCP Apps समर्थन घोषित नहीं करता है और उसी कनेक्टर पर इसका उपयोग करने से Chat में इनलाइन ऐप्स गायब हो सकते हैं, जब तक आप कनेक्टर - Access to secretsline 269
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/mcp-clients.mdx
- Network accessline 148
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 149
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 178
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 179
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 180
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 226
[Credential-related source omitted] - Network accessline 274
Atlassian प्रीसेट, Atlassian Rovo MCP के OAuth 2.1 एंडपॉइंट `https://mcp.atlassian.com/v1/mcp/authv2` का उपयोग करता है और Jira तथा Confluence टूल उपलब्ध कराता है। Jira टिकट बनाने के लिए Atlassian की वे अनुमतियां आवश्यक हैं जो Jira में लिखने - Network accessline 275
`https://mcp.atlassian.com/v1/mcp/authv2` का उपयोग करता है और Jira तथा Confluence टूल उपलब्ध कराता है। Jira टिकट बनाने के लिए Atlassian की वे अनुमतियां आवश्यक हैं जो Jira में लिखने की अनुमति देती हैं; उपयोगकर्ता को फिर भी केवल वही प्रोजेक्ट - Network accessline 277
Atlassian संगठन के व्यवस्थापक नियंत्रित कर सकते हैं कि किसी OAuth क्लाइंट को कनेक्ट करने की अनुमति है या नहीं। कनेक्ट करने से पहले, किसी व्यवस्थापक को ऐप का वास्तविक HTTPS ओरिजिन **Atlassian Administration → Rovo → Rovo MCP server → Domains - Network accessline 283
1. **सर्वर नाम** — आपके स्वयं के संदर्भ के लिए एक संक्षिप्त लेबल (उदाहरण के लिए "ज़ैपियर", "स्लैक-टूल्स")। 2. **URL** - MCP सर्वर प्रदाता द्वारा आपको दिया गया HTTPS एंडपॉइंट (उदाहरण के लिए `https://mcp.zapier.com/s/abc123/mcp`)। यह आमतौर पर - Network accessline 284
2. **URL** - MCP सर्वर प्रदाता द्वारा आपको दिया गया HTTPS एंडपॉइंट (उदाहरण के लिए `https://mcp.zapier.com/s/abc123/mcp`)। यह आमतौर पर प्रदाता के डैशबोर्ड या एकीकरण दस्तावेज़ों में पाया जाता है। - Network accessline 299
[Credential-related source omitted] - Network accessline 391
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 392
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 393
"url": "https://example.com/mcp", - Network accessline 431
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com
packages/core/docs/content/locales/hi-IN/mcp-protocol.mdx
- Access to secretsline 31
[Credential-related source omitted] - Network accessline 108
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 117
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 118
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 119
"url": "https://mail.example.com/mcp", - Network accessline 138
POST https://your-app.example.com/mcp - Access to secretsline 143
[Credential-related source omitted] - Access to secretsline 262
[Credential-related source omitted] - Network accessline 270
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 276
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 277
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 392
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 401
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 402
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 403
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/locales/hi-IN/messaging.mdx
- Network accessline 81
1. **[api.slack.com/apps](https://api.slack.com/apps)** पर जाएं और **नया ऐप बनाएं** → **स्क्रैच से** पर क्लिक करें। इसे नाम दें (उदाहरण के लिए "एजेंट") और अपना कार्यक्षेत्र चुनें। - Network accessline 96
```पाठ https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 97
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 120
https://your-clips.example.com/api/slack/unfurl - Access to secretsline 134
[Credential-related source omitted] - Network accessline 167
OAuth Redirect URL को `https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 168
`https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 169
Event Subscriptions Request URL को `https://your-app.example.com/_agent-native/integrations/slack/webhook`, और Interactivity Request URL को - Network accessline 170
`https://your-app.example.com/_agent-native/integrations/slack/webhook`, और - Network accessline 171
Interactivity Request URL को `https://your-app.example.com/_agent-native/integrations/slack/interactions` पर सेट करें। `app_home_opened`, `app_context_changed`, `app_mention` और `message.im` की सदस्यता लें। - Network accessline 172
`https://your-app.example.com/_agent-native/integrations/slack/interactions` पर सेट करें। - Access to secretsline 174
[Credential-related source omitted] - Network accessline 262
[Credential-related source omitted] - Access to secretsline 262
[Credential-related source omitted] - Access to secretsline 267
[Credential-related source omitted] - Access to secretsline 364
[Credential-related source omitted] - Network accessline 378
1. टेलीग्राम खोलें और **[@BotFather](https://t.me/BotFather)** संदेश भेजें। - Network accessline 385
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 404
1. **[resend.com](https://resend.com)** पर साइन अप करें। आरंभ करने के लिए निःशुल्क टियर पर्याप्त है। - Network accessline 411
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 418
[Credential-related source omitted] - Network accessline 425
1. **[sendgrid.com](https://sendgrid.com)** पर साइन अप करें। - Network accessline 433
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 438
[Credential-related source omitted] - Network accessline 462
1. **[Meta Developer Portal](https://developers.facebook.com/)** पर जाएं, **ऐप बनाएं** पर क्लिक करें, और **व्यवसाय** प्रकार चुनें। - Access to secretsline 469
[Credential-related source omitted] - Network accessline 475
कॉलबैक URL: https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Access to secretsline 686
[Credential-related source omitted] - Access to secretsline 687
[Credential-related source omitted] - Access to secretsline 797
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/multi-app-workspace.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 100
[Credential-related source omitted] - Network accessline 245
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 246
"urls": { "design": "https://design.example.test" } - Network accessline 247
"design": "https://design.example.test" - Access to secretsline 276
[Credential-related source omitted] - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Network accessline 325
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 326
# https://your-agents.com/mail/* → apps/mail - Network accessline 327
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 328
# https://your-agents.com/forms/* → apps/forms
packages/core/docs/content/locales/hi-IN/native-chat-ui.mdx
- Network accessline 449
पुराने और नए API का मानचित्र और माइग्रेशन चरण देखने के लिए [AgentKit चैट माइग्रेशन गाइड](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md) देखें।
packages/core/docs/content/locales/hi-IN/netlify.mdx
packages/core/docs/content/locales/hi-IN/node-js.mdx
packages/core/docs/content/locales/hi-IN/notifications.mdx
- Network accessline 192
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 193
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 194
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 211
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 212
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 213
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/locales/hi-IN/observability.mdx
- Network accessline 507
रिपॉज़िटरी के [`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel) में इस प्लगइन का अधिक पूर्ण संस्करण है, जिसमें हर सिग्नल के लिए अलग एंडपॉइंट और एक्सपोर्टर सत्यापन है। यह npm पर प्रकाशित नहीं है, इसलिए इसे
packages/core/docs/content/locales/hi-IN/onboarding.mdx
- Access to secretsline 187
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/plan-plugin.mdx
- Network accessline 21
- **तीन कौशल** — `/visual-plan` (कैनॉनिकल एंट्री पॉइंट), `/visual-recap`, और `/visualize-repo`। - **Plan MCP कनेक्टर** — `https://plan.agent-native.com` पर होस्टेड ऐप के - Network accessline 22
`/visualize-repo`। - **Plan MCP कनेक्टर** — `https://plan.agent-native.com` पर होस्टेड ऐप के विरुद्ध रजिस्टर्ड (MCP एंडपॉइंट `https://plan.agent-native.com/mcp`, सर्वर - Network accessline 23
- **Plan MCP कनेक्टर** — `https://plan.agent-native.com` पर होस्टेड ऐप के - Network accessline 24
विरुद्ध रजिस्टर्ड (MCP एंडपॉइंट `https://plan.agent-native.com/mcp`, सर्वर - Network accessline 102
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 129
- **Chromium ब्राउज़र का उपयोग करें।** Safari होस्टेड HTTPS Plan पेज को `http://127.0.0.1` लोकलहोस्ट ब्रिज पढ़ने से रोकता है (मिक्स्ड-कंटेंट / - Network accessline 130
`http://127.0.0.1` लोकलहोस्ट ब्रिज पढ़ने से रोकता है (मिक्स्ड-कंटेंट / - Network accessline 238
लिए MCP कॉन्फ़िग लिखनी है (डिफ़ॉल्ट `all`)। - `--no-connect` — प्रमाणीकरण के बिना कनेक्टर रजिस्टर करें; बाद में `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 239
- `--no-connect` — प्रमाणीकरण के बिना कनेक्टर रजिस्टर करें; बाद में `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` चलाएं, या एक ज़्यादा सीमित `--client` चुनें। - Network accessline 240
`npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 303
इंस्टॉल के बाद, **एक नया Codex थ्रेड शुरू करें** ताकि कौशल और MCP टूल सेशन में लोड हों। प्लगइन एक URL-only कनेक्टर शिप करता है (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`); `codex mcp login plan` OAuth फ़्लो - Network accessline 304
लोड हों। प्लगइन एक URL-only कनेक्टर शिप करता है (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`); `codex mcp login plan` OAuth फ़्लो चलाता है। ऊपर दिया गया यूनिवर्सल CLI रूट Codex के लिए भी काम करता है - Network accessline 305
`https://plan.agent-native.com/mcp`); `codex mcp login plan` OAuth फ़्लो - Network accessline 313
**पुराने इंस्टॉल:** अगर आपके कॉन्फ़िग में अभी भी वही URL पॉइंट करने वाली एक `agent-native-plans` एंट्री है, तो Codex के लिए `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 314
`agent-native-plans` एंट्री है, तो Codex के लिए `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` चलाना, या अपने टारगेट `--client` के साथ वही कमांड चलाना, इसे कैनॉनिकल `plan` - Network accessline 315
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`
packages/core/docs/content/locales/hi-IN/pr-visual-recap.mdx
- Access to secretsline 217
[Credential-related source omitted] - Access to secretsline 218
[Credential-related source omitted] - Access to secretsline 219
[Credential-related source omitted] - Network accessline 251
DeepSeek के लिए, `VISUAL_RECAP_AGENT=openai-compatible`, `VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`, - Network accessline 252
`VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`, - Access to secretsline 253
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Access to secretsline 311
[Credential-related source omitted] - Network accessline 324
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 340
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 352
[Credential-related source omitted] - Access to secretsline 353
[Credential-related source omitted] - Network accessline 359
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | वेरिएबल, सिर्फ़ कॉपी वेरिएंट के लिए (रीयूज़ेबल caller इसे अनदेखा करता है)। `false` पर सेट करने पर हर पुश पर रीकैप बनना बंद हो जाता है। इस - Network accessline 361
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | सीक्रेट। सिर्फ़ तभी जब Plans ऐप को किसी अलग ओरिजिन पर सेल्फ-होस्ट किया जाए। - Network accessline 406
इसे इंस्टॉल करने के लिए, फ़ाइल को [BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Network accessline 407
[BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Access to secretsline 410
[Credential-related source omitted] - Access to secretsline 638
[Credential-related source omitted] - Access to secretsline 706
[Credential-related source omitted] - Access to secretsline 707
[Credential-related source omitted] - Access to secretsline 708
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/privacy-and-data-rights.mdx
- Network accessline 59
- [GDPR पाठ](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [European Commission: individual requests](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 60
- [European Commission: individual requests](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [California Attorney General: CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [CCPA regulations](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/locales/hi-IN/real-time-collaboration.mdx
- Network accessline 681
`GET /_agent-native/health` परिणाम बताता है, और उसे हल करने की प्रक्रिया ही पंजीकरण कर देती है, इसलिए एक ही curl से सक्रियण और पुष्टि दोनों हो जाते हैं: - Network accessline 682
पंजीकरण कर देती है, इसलिए एक ही curl से सक्रियण और पुष्टि दोनों हो जाते हैं: - Network accessline 685
curl -s https://your-app.example.com/_agent-native/health | jq .realtime
packages/core/docs/content/locales/hi-IN/recurring-jobs.mdx
packages/core/docs/content/locales/hi-IN/render.mdx
packages/core/docs/content/locales/hi-IN/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- कभी भी `innerHTML`, `eval()`, या `document.write()` का उपयोग न करें - Access to secretsline 393
[Credential-related source omitted] - Access to secretsline 409
[Credential-related source omitted] - Access to secretsline 415
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/server-database.mdx
- Network accessline 8
हर एजेंट-नेटिव ऐप अपनी स्थिति PostgreSQL में संग्रहीत करता है। UI और एजेंट दोनों एक ही [Drizzle ORM](https://orm.drizzle.team) क्लाइंट का उपयोग करते हुए, एक ही Actions के माध्यम से एक ही तालिकाओं को पढ़ते और लिखते हैं। लाइव सिंक SSE का उपयो - Network accessline 81
[PGlite](https://pglite.dev/) एक इन-प्रोसेस WASM Postgres डेटाबेस चलाता है। यह आपको तैनाती से पहले Postgres-केवल स्कीमा समस्याओं को पकड़ने देता है, जबकि सेटअप को PGlite डिफ़ॉल्ट जितना ही सरल रखता है। PGlite की तरह, यह भी केवल-स्थानीय संग्रह
packages/core/docs/content/locales/hi-IN/server-middleware.mdx
- Network accessline 104
अधिक मिडलवेयर पैटर्न - सशर्त लॉजिक, असिंक्रोनस ऑपरेशन, प्रतिक्रिया इंटरसेप्शन - के लिए [Nitro मिडलवेयर डॉक्स](https://nitro.build/guide/routing#middleware) और [H3 डॉक्स](https://h3.unjs.io) देखें। - Network accessline 147
- [**रूट्स**](/docs/server-routes): कस्टम रूट्स जो इस मिडलवेयर द्वारा सुरक्षित हैं - [**Security**](/docs/security): डेटा स्कोपिंग, access गार्ड, और पूरा auth मॉडल - [**Nitro मिडलवेयर डॉक्स**](https://nitro.build/guide/routing#middleware): - Network accessline 149
- [**Nitro मिडलवेयर डॉक्स**](https://nitro.build/guide/routing#middleware): पूरा मिडलवेयर API और पैटर्न
packages/core/docs/content/locales/hi-IN/server-overview.mdx
- Network accessline 204
[Nitro](https://nitro.build) वह सर्वर टूलकिट है जो एजेंट-नेटिव HTTP परत को शक्ति देता है। `server/` में आप जो अधिकांश चीज़ें कॉन्फ़िगर करते हैं वह Nitro ही है: फ़ाइल-आधारित रूट्स, स्टार्टअप प्लगइन्स, और मिडलवेयर। - Network accessline 212
अंतर्निहित अनुरोध/प्रतिक्रिया API [H3](https://h3.unjs.io) है। जब आप कस्टम रूट्स में `defineEventHandler`, `createError`, या `getQuery` देखते हैं, वह H3 है। [Nitro डॉक्स](https://nitro.build/guide) और [H3 डॉक्स](https://h3.unjs.io) उस किसी
packages/core/docs/content/locales/hi-IN/server-plugins.mdx
- Network accessline 139
प्लगइन्स Nitro के किसी भी लाइफ़साइकिल हुक का उपयोग कर सकते हैं: `request`, `beforeResponse`, `close`, और अन्य। [Nitro प्लगइन डॉक्स](https://nitro.build/guide/plugins) पूरी हुक API और प्लगइन अनुबंध को विस्तार से कवर करते हैं। - Network accessline 161
- [**रूट्स**](/docs/server-routes): कस्टम फ़ाइल रूट्स और उनका उपयोग कब करें - [**लेखन एजेंट निर्देश और कौशल**](/docs/writing-agent-instructions): agent-chat प्लगइन में `systemPrompt` कैसे तैयार करें - [**Nitro प्लगइन डॉक्स**](https://nitro. - Network accessline 163
- [**Nitro प्लगइन डॉक्स**](https://nitro.build/guide/plugins): पूरी प्लगइन API और लाइफ़साइकिल हुक
packages/core/docs/content/locales/hi-IN/server-routes.mdx
- Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/skills-guide.mdx
- Network accessline 315
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 316
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 317
"url": "https://assets.agent-native.com", - Network accessline 318
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/locales/hi-IN/template-analytics.mdx
packages/core/docs/content/locales/hi-IN/template-assets.mdx
- Network accessline 115
लाइव डेमो: [assets.agent-native.com](https://assets.agent-native.com). - Network accessline 464
<EmbeddedApp url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 465
url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 501
एसेट ऐप स्किल में ऐप आईडी `assets` और होस्ट किया गया MCP URL है `https://assets.agent-native.com/mcp`. - Network accessline 502
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/locales/hi-IN/template-brain.mdx
- Network accessline 159
लाइव डेमो: [brain.agent-native.com](https://brain.agent-native.com).
packages/core/docs/content/locales/hi-IN/template-calendar-developers.mdx
- Network accessline 30
इसके लिए पहले [Node.js](https://nodejs.org) 22.22 या उसके बाद का वर्शन और - Network accessline 31
[pnpm](https://pnpm.io) इंस्टॉल होना चाहिए। अगर आपने अभी तक इनमें से कोई - Network accessline 42
ऐप `http://localhost:8082` पर खुलता है। अगर यह अपने आप न खुले, तो अपने - Network accessline 53
1. **एक Google Cloud प्रोजेक्ट बनाएं**, अगर आपके पास इस्तेमाल करने के लिए पहले से कोई नहीं है: [Google Cloud Console](https://console.cloud.google.com/) - Network accessline 54
पहले से कोई नहीं है: [Google Cloud Console](https://console.cloud.google.com/) - Network accessline 64
4. **क्रेडेंशियल बनाएं**: **APIs & Services → Credentials** के तहत, "Web application" टाइप का एक OAuth client ID बनाएं, और `http://localhost:8082/_agent-native/google/callback` को authorized - Network accessline 65
application" टाइप का एक OAuth client ID बनाएं, और `http://localhost:8082/_agent-native/google/callback` को authorized redirect URI के तौर पर जोड़ें (8082 कैलेंडर का डिफ़ॉल्ट डेव पोर्ट है)। - Network accessline 66
`http://localhost:8082/_agent-native/google/callback` को authorized - Access to secretsline 83
[Credential-related source omitted] - Network accessline 86
[Zoom OAuth app](https://marketplace.zoom.us/) से एक Client ID और Secret लें
packages/core/docs/content/locales/hi-IN/template-calendar-features.mdx
packages/core/docs/content/locales/hi-IN/template-calendar-integrations.mdx
- Access to secretsline 24
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/template-calendar.mdx
packages/core/docs/content/locales/hi-IN/template-chat-developers.mdx
- Network accessline 30
इसके लिए [Node.js](https://nodejs.org) 22.22 या उसके बाद का वर्शन चाहिए। अगर `pnpm`
packages/core/docs/content/locales/hi-IN/template-chat-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/template-chat.mdx
packages/core/docs/content/locales/hi-IN/template-clips-developers.mdx
- Network accessline 24
इसके लिए पहले [Node.js](https://nodejs.org) 22.22 या उसके बाद का वर्शन और - Network accessline 25
[pnpm](https://pnpm.io) इंस्टॉल होना चाहिए। अगर आपने अभी तक इनमें से कोई - Network accessline 330
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 331
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 332
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/template-clips-embed.mdx
- Network accessline 60
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 81
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 82
[Credential-related source omitted] - Network accessline 84
const response = await fetch( - Network accessline 206
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 214
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/locales/hi-IN/template-clips-features.mdx
- Network accessline 185
कैमरा बबल के साथ स्क्रीन और वेबकैम रिकॉर्डिंग, रिकॉर्ड करते समय लाइव ट्रांसक्रिप्शन, यह सब बिना किसी ब्राउज़र टैब को खुला रखे। इसे [clips.agent-native.com/download](https://clips.agent-native.com/download) से - Network accessline 186
ट्रांसक्रिप्शन, यह सब बिना किसी ब्राउज़र टैब को खुला रखे। इसे [clips.agent-native.com/download](https://clips.agent-native.com/download) से डाउनलोड करें, जो macOS, Windows, और Linux के लिए उपलब्ध है। - Network accessline 187
[clips.agent-native.com/download](https://clips.agent-native.com/download) से - Network accessline 319
[Credential-related source omitted] - Network accessline 320
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/template-clips-integrations.mdx
- Access to secretsline 24
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/template-clips.mdx
packages/core/docs/content/locales/hi-IN/template-content.mdx
- Network accessline 17
### [अभी आज़माएं →](https://content.agent-native.com)
packages/core/docs/content/locales/hi-IN/template-design-developers.mdx
packages/core/docs/content/locales/hi-IN/template-design-integrations.mdx
- Access to secretsline 25
[Credential-related source omitted] - Network accessline 51
https://design.agent-native.com/mcp - Network accessline 56
फिर ऑथेंटिकेट करने के लिए `/mcp` चलाता है। Codex और दूसरे CLI या stdio क्लाइंट इसके बजाय `npx @agent-native/core@latest connect https://design.agent-native.com` चलाते हैं, जो अधिकृत करने के लिए एक - Network accessline 57
क्लाइंट इसके बजाय `npx @agent-native/core@latest connect https://design.agent-native.com` चलाते हैं, जो अधिकृत करने के लिए एक ब्राउज़र टैब खोलता है और बिना किसी शेयर्ड सीक्रेट के एक पर-यूज़र, - Network accessline 58
https://design.agent-native.com` चलाते हैं, जो अधिकृत करने के लिए एक - Network accessline 73
लौटाने लगे, तो उसे बार-बार रिट्राई न करें। Claude Code में `/mcp` चलाएं और दोबारा कनेक्ट करें, या किसी भी टर्मिनल से `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` - Network accessline 74
और दोबारा कनेक्ट करें, या किसी भी टर्मिनल से `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` चलाएं। दोनों बिना कुछ भी दोबारा इंस्टॉल किए दोबारा ऑथेंटिकेट कर देते हैं। - Network accessline 75
@agent-native/core@latest reconnect https://design.agent-native.com`
packages/core/docs/content/locales/hi-IN/template-design.mdx
packages/core/docs/content/locales/hi-IN/template-dispatch.mdx
packages/core/docs/content/locales/hi-IN/template-forms-agent.mdx
- Network accessline 40
- **"इसे प्रकाशित करें और मुझे लिंक दें।"** फॉर्म लाइव हो जाता है और आपको तुरंत इसका सार्वजनिक URL मिल जाता है। - **"एक गुमनाम कस्टमर फ़ीडबैक फॉर्म बनाएं और मुझे सार्वजनिक लिंक दें।"** फॉर्म शुरू से ही गुमनाम प्रकाशित होता है, इसलिए सबमिट क - Network accessline 42
- **"इस फॉर्म को सिर्फ़ https://example.com से ही सबमिट होने दें।"** फॉर्म कहीं और से आने वाले सबमिशन रिजेक्ट करना शुरू कर देता है। आज यह सिर्फ़ एजेंट से पूछकर ही उपलब्ध है। बिल्डर में इसके लिए कोई टॉगल नहीं है।
packages/core/docs/content/locales/hi-IN/template-forms-features.mdx
- Network accessline 14
[Credential-related source omitted] - Network accessline 15
[Credential-related source omitted] - Network accessline 94
सार्वजनिक सबमिशन के लिए, आगे बढ़ने से पहले [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/) के ज़रिए सत्यापन ज़रूरी बनाया जा सकता है। Settings में एक साइट की और एक सीक्रेट की सेट करके इसे चालू करें। Turnstile के बिना भ - Network accessline 110
आज बिल्डर में इसके लिए कोई टॉगल नहीं है। एजेंट से इसे सेट करने के लिए कहें, उदाहरण के लिए "इस फॉर्म को सिर्फ़ https://example.com से सबमिट होने दें।" जिस भी साइट को अनुमति देनी हो, उसके लिए एजेंट को सटीक ऑरिजिन दें (स्कीम और डोमेन, जैसे `ht - Network accessline 116
[Credential-related source omitted] - Network accessline 117
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/template-forms.mdx
packages/core/docs/content/locales/hi-IN/template-mail.mdx
- Network accessline 101
लाइव डेमो: [mail.agent-native.com](https://mail.agent-native.com). - Network accessline 172
1. [Google Cloud Console](https://console.cloud.google.com/) खोलें और एक प्रोजेक्ट बनाएं। - Network accessline 173
2. APIs और सेवाओं → लाइब्रेरी के अंतर्गत **Gmail API** को सक्षम करें। 3. OAuth 2.0 क्रेडेंशियल बनाएं (प्रकार: वेब एप्लिकेशन)। `http://localhost:8085/_agent-native/google/callback` को अधिकृत रीडायरेक्ट URI के रूप में जोड़ें। 4. क्लाइंट आईडी - Network accessline 174
3. OAuth 2.0 क्रेडेंशियल बनाएं (प्रकार: वेब एप्लिकेशन)। `http://localhost:8085/_agent-native/google/callback` को अधिकृत रीडायरेक्ट URI के रूप में जोड़ें।
packages/core/docs/content/locales/hi-IN/template-plan.mdx
- Network accessline 138
यह तीन स्लैश कमांड इंस्टॉल करता है — `/visual-plan`, `/visual-recap`, और `/visualize-repo` — साथ ही `https://plan.agent-native.com/mcp` पर `plan` MCP - Network accessline 139
`/visualize-repo` — साथ ही `https://plan.agent-native.com/mcp` पर `plan` MCP - Network accessline 236
अगर कोई Plans टूल कभी `needs auth`, `Unauthorized`, या `Session terminated` लौटाता है, तो उसे दोबारा कोशिश करते न रहें। `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 237
लौटाता है, तो उसे दोबारा कोशिश करते न रहें। `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` के साथ फिर से प्रमाणित करें, या किसी OAuth-सक्षम होस्ट के लिए फिर से `/mcp` - Network accessline 238
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`
packages/core/docs/content/locales/hi-IN/template-slides-features.mdx
packages/core/docs/content/locales/hi-IN/template-slides-integrations.mdx
- Access to secretsline 24
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/template-slides.mdx
packages/core/docs/content/locales/hi-IN/toolkit-capability-packages.mdx
packages/core/docs/content/locales/hi-IN/toolkit-settings.mdx
- Access to secretsline 139
[Credential-related source omitted] - Network accessline 230
इस इम्प्लीमेंटेशन के लिए OAuth callback उपयोगकर्ता की मशीन के loopback पते `127.0.0.1` पर होना चाहिए; `localhost` इसका स्वीकार्य विकल्प नहीं है। यह स्वयं-सेवा फ़्लो होस्टेड `*.agent-native.com` साइटों पर समर्थित नहीं है। पार्टनर मंज़ूरी होस - Network accessline 240
Inference के लिए केवल सार्वजनिक Responses API इस्तेमाल करें। ChatGPT के निजी backend endpoints और ChatGPT बातचीत का डेटा इस अनुमति के दायरे में नहीं आते। OpenAI टोकन साझा करने की प्रक्रिया को पूर्वावलोकन बताता है, इसलिए मॉडल की उपलब्धता, सी
packages/core/docs/content/locales/hi-IN/tracking.mdx
- Network accessline 83
| प्रदाता | पर्यावरण संस्करण | | --------- | ------------------------------------------------------------------------------------------ | | पोस्टहॉग | `POSTHOG_API - Network accessline 84
| --------- | ------------------------------------------------------------------------------------------ | | पोस्टहॉग | `POSTHOG_API_KEY` (आवश्यक), `POSTHOG_HOST` (वैकल्पिक, डिफ़ॉल्ट `https://us.i.posthog.com`) | | मिक्सपैनल | `MIXPANEL_TO - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Network accessline 205
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 206
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 207
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 225
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 226
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 283
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 284
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 285
fetch("https://analytics.example.com/events", { - Network accessline 343
- **बैच HTTP** - अंतर्निहित प्रदाता घटनाओं को सूचीबद्ध करते हैं और हर 10 सेकंड में या जब 50 घटनाएं जमा हो जाती हैं, जो भी पहले हो, फ्लश करते हैं। यह डेटा खोए बिना आउटबाउंड अनुरोधों को न्यूनतम करता है। - **कोई SDK निर्भरता नहीं** - सभी अंतर् - Network accessline 344
- **कोई SDK निर्भरता नहीं** - सभी अंतर्निहित प्रदाता कच्चे `fetch()` का उपयोग करते हैं। कोई पोस्टहॉग SDK नहीं, कोई मिक्सपैनल SDK नहीं, कोई आयाम SDK नहीं। फ्रेमवर्क को हल्का रखता है।
packages/core/docs/content/locales/hi-IN/turn-into-app.mdx
- Network accessline 12
- एक AI एजेंट जो टर्मिनल कमांड चला सके, जैसे Claude Code, Codex या Cursor। वेब पर Claude या ChatGPT इसके बजाय काम Builder को सौंप सकते हैं। देखें [यह कहाँ चलती है](#hosts)। - [Node.js](https://nodejs.org)। nodejs.org का इंस्टॉलर `npx` के सा - Network accessline 13
- [Node.js](https://nodejs.org)। nodejs.org का इंस्टॉलर `npx` के साथ आता है, जो नीचे दिए इंस्टॉल कमांड चलाता है। - Network accessline 18
- **आपकी मशीन पर।** Node.js 22.22 या उससे नया (`node --version` से जाँचें) और आपके `PATH` में [pnpm](https://pnpm.io)। यदि `pnpm --version` विफल हो, तो `corepack enable` चलाएँ, जो Node.js के साथ आता है। - Network accessline 59
Builder होस्टिंग के लिए, किसी भी तरह के होस्ट से, Agent-Native Dispatch MCP कनेक्टर (`https://dispatch.agent-native.com/mcp`, OAuth) चाहिए और स्किल उसका `start-workspace-app-creation` एक्शन कॉल करती है। आपकी ओर कुछ नहीं चलता, इसलिए Dispatch - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/updating-ui-in-production.mdx
- Network accessline 18
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment) एक अलग managed code-capable workflow है जहां एक agent production में आपके ऐप का UI modify कर सकता है। अपने repo के सा
packages/core/docs/content/locales/hi-IN/upgrading-core-ui.mdx
- Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 74
[Credential-related source omitted] - Network accessline 183
Canonical source URL: [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx). - Network accessline 184
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx).
packages/core/docs/content/locales/hi-IN/vercel.mdx
- Access to secretsline 96
[Credential-related source omitted] - Network accessline 113
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 140
[Credential-related source omitted] - Network accessline 203
Scheduled automations, queued event automations, और बाधित runs की सफ़ाई, ये सब हर मिनट एक बार चलने वाले sweep पर होते हैं। Build एक [Vercel cron job](https://vercel.com/docs/cron-jobs) जोड़ता है जो - Network accessline 204
हर मिनट एक बार चलने वाले sweep पर होते हैं। Build एक [Vercel cron job](https://vercel.com/docs/cron-jobs) जोड़ता है जो `/_agent-native/jobs/_process-sweep` पर sweep को call करता है। कार्यस्थान build - Network accessline 205
[Vercel cron job](https://vercel.com/docs/cron-jobs) जोड़ता है जो - Access to secretsline 209
[Credential-related source omitted]
packages/core/docs/content/locales/hi-IN/voice-input.mdx
packages/core/docs/content/locales/hi-IN/what-is-agent-native.mdx
packages/core/docs/content/locales/hi-IN/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Access to secretsline 124
[Credential-related source omitted] - Network accessline 267
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 268
[Credential-related source omitted] - Network accessline 269
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 270
[Credential-related source omitted] - Access to secretsline 271
[Credential-related source omitted] - Access to secretsline 272
[Credential-related source omitted] - Network accessline 284
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 285
[Credential-related source omitted] - Network accessline 286
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 287
[Credential-related source omitted] - Access to secretsline 288
[Credential-related source omitted] - Access to secretsline 289
[Credential-related source omitted] - Network accessline 340
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 342
APP_URL=https://workspace.example.com
packages/core/docs/content/locales/hi-IN/workspace-management.mdx
- Network accessline 90
हर कोई जिसे परिवर्तन करने की आवश्यकता है, वह गिट के साथ सहज नहीं है। [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management) एक विज़ुअल ब्रांचिंग मॉडल का समर्थन कर - Network accessline 140
मुख्य युक्तियाँ: GitHub टीमों (`@org/team`) का उपयोग करें, व्यक्तियों का नहीं। फ़्रेमवर्क और डिस्पैच परिवर्तनों के लिए हमेशा प्लेटफ़ॉर्म समीक्षा की आवश्यकता होनी चाहिए। ग्लोब सिंटैक्स और मल्टीपल-ओनर पैटर्न के लिए [GitHub CODEOWNERS docs](ht - Network accessline 160
फिर [actions/labeler](https://github.com/actions/labeler) क्रिया जोड़ें - संपूर्ण वर्कफ़्लो YAML के लिए उस रेपो का README देखें। जब पीआर खोले या अपडेट किए जाते हैं तो लेबल स्वचालित रूप से लागू हो जाते हैं। - Access to secretsline 233
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/a2a-protocol.mdx
- Access to secretsline 93
[Credential-related source omitted] - Network accessline 117
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 118
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 119
"url": "https://analytics.example.com", - Network accessline 314
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 350
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 351
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 352
"https://analytics.example.com", - Access to secretsline 354
[Credential-related source omitted] - Network accessline 374
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", - Network accessline 375
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", "cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 376
"url": "https://example.services.ai.azure.com/a2a", - Network accessline 377
"cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 417
[Azure AI Foundry のホスト型エージェント](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 424
Foundry は [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint) - Network accessline 435
[Gemini Enterprise は外部 A2A エージェントを登録できます](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent)。 - Network accessline 439
`POST {url}/v1/message:stream` を使います。リクエストには Google OAuth または ADC bearer トークンと Google Cloud IAM 権限が必要です。詳しくは [Registry A2A エンドポイントでエージェントを呼び出す](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) - Network accessline 440
または ADC bearer トークンと Google Cloud IAM 権限が必要です。詳しくは [Registry A2A エンドポイントでエージェントを呼び出す](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) を参照してください。 - Network accessline 441
[Registry A2A エンドポイントでエージェントを呼び出す](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) - Network accessline 459
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 460
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 461
"url": "https://api.anthropic.com", - Access to secretsline 466
[Credential-related source omitted] - Access to secretsline 479
[Credential-related source omitted] - Network accessline 525
# デプロイ済みアプリを URL で直接呼び出し、環境変数からベアラー トークンを読み取る agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 526
[Credential-related source omitted] - Access to secretsline 526
[Credential-related source omitted] - Access to secretsline 527
[Credential-related source omitted] - Access to secretsline 599
[Credential-related source omitted] - Access to secretsline 603
[Credential-related source omitted] - Network accessline 661
- 共有 `A2A_SECRET` は、ビルド時にすべてのアプリの環境にマウントされます。 - クロスアプリ呼び出しは同じ発信元 (`https://workspace.example.com/apps/analytics` が `https://workspace.example.com/apps/mail` を呼び出す) であるため、DNS、CORS、またはペアごとの JWT セットアップはありません。 - Network accessline 662
- クロスアプリ呼び出しは同じ発信元 (`https://workspace.example.com/apps/analytics` が `https://workspace.example.com/apps/mail` を呼び出す) であるため、DNS、CORS、またはペアごとの JWT セットアップはありません。 - Network accessline 670
**応答を超えて存続するファイアアンドフォーゲット `Promise` には決して依存しないでください。** サーバーレス関数 (Netlify、Vercel、AWS Lambda、Cloud Run) は、応答本文がフラッシュされた瞬間にフリーズします。場合によっては、待機していない `fetch(...)` の TCP ハンドシェイクが完了する前にフリーズします。ノード上でローカルに動作するパターンは、実稼働環境での作業を黙って削除します。 - Network accessline 708
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 709
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 710
"https://analytics.example.com", - Access to secretsline 712
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/actions-defining.mdx
- Network accessline 226
`schema` は、[Standard Schema](https://standardschema.dev) 互換のライブラリであれば何でも受け付けます。
packages/core/docs/content/locales/ja-JP/add-ui-to-your-agent.mdx
packages/core/docs/content/locales/ja-JP/agent-mentions.mdx
packages/core/docs/content/locales/ja-JP/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/agent-web-surfaces.mdx
- Network accessline 171
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 172
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 173
siteUrl: "https://example.com", - Network accessline 198
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/locales/ja-JP/audit-log.mdx
- Network accessline 110
**`export-audit-ocsf`** は、組織の監査証跡を [OCSF](https://schema.ocsf.io) 1.9.0 の API Activity イベント (`class_uid` 6003) として古い順に返します。SIEM が定期的に取得するためのものです。オーナーと管理者のみが利用でき、組織の監査ログと同じ `org` と `admins` のイベントを読み取ります。誰かのプライベートな行は読み取りません。
packages/core/docs/content/locales/ja-JP/authentication.mdx
- Network accessline 8
エージェント ネイティブ アプリは、アカウント優先の設計で認証に [Better Auth](https://better-auth.com) を使用します。ユーザーは初回訪問時にアカウントを作成し、初日から実際の ID を取得します。 - Access to secretsline 194
[Credential-related source omitted] - Access to secretsline 199
[Credential-related source omitted] - Access to secretsline 203
[Credential-related source omitted] - Access to secretsline 208
[Credential-related source omitted] - Access to secretsline 210
[Credential-related source omitted] - Access to secretsline 232
[Credential-related source omitted] - Access to secretsline 236
[Credential-related source omitted] - Network accessline 249
https://mail.agent-native.com/mcp - Access to secretsline 474
[Credential-related source omitted] - Access to secretsline 476
[Credential-related source omitted] - Access to secretsline 478
[Credential-related source omitted] - Access to secretsline 479
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/automations.mdx
packages/core/docs/content/locales/ja-JP/aws-amplify.mdx
- Access to secretsline 87
[Credential-related source omitted] - Network accessline 99
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 154
[Credential-related source omitted] - Network accessline 168
[Amplify コンソール](https://console.aws.amazon.com/amplify/)でアプリを作成します。 - Network accessline 196
最初のビルドには数分かかります。その進行状況は Amplify コンソールで確認できます。完了したら、`https://main.d1m7bkiki6tdw1.amplifyapp.com` のような形式のアプリのアドレスを開き、最初のアカウントを作成します。 - Access to secretsline 233
[Credential-related source omitted] - Access to secretsline 237
[Credential-related source omitted] - Access to secretsline 242
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/aws-lambda.mdx
- Network accessline 31
AWS アカウントと、自分のコンピューターにインストールしてサインイン済みの [AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html) も必要です。AWS CLI にはデフォルトのリージョンを設定しておいてください。このガイドのコマンドは自分のコンピューター上で実行してください。ステップ 5 と 6 では AWS CLI を使い、ステップ 7 では Clou - Access to secretsline 90
[Credential-related source omitted] - Access to secretsline 116
[Credential-related source omitted] - Access to secretsline 133
[Credential-related source omitted] - Access to secretsline 176
[Credential-related source omitted] - Access to secretsline 184
[Credential-related source omitted] - Network accessline 203
最初のコマンドは `FunctionUrl` を出力します。これは `https://abc123.lambda-url.us-east-1.on.aws/` のような形式です。ステップ 7 のためにコピーしておいてください。2 つの `add-permission` コマンドは、そのアドレス経由でリクエストが関数に届くことを許可します。 - Network accessline 219
[CloudFront コンソール](https://console.aws.amazon.com/cloudfront/v4/home)で CloudFront ディストリビューションを作成します。 - Network accessline 229
ディストリビューションの **Origins** タブで、オリジンを作成します。**Origin domain** には、ステップ 5 の関数 URL のホスト名を、`https://` と末尾のスラッシュを除いて貼り付けます。**Protocol** を **HTTPS only** に、**Response timeout** を 60 秒に設定します。このオリジンにはオリジン アクセス コントロールを追加しないでください。 - Network accessline 252
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 254
[Credential-related source omitted] - Access to secretsline 256
[Credential-related source omitted] - Access to secretsline 270
[Credential-related source omitted] - Dangerous shell operationline 272
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/locales/ja-JP/azure-static-web-apps.mdx
packages/core/docs/content/locales/ja-JP/blueprint-installer.mdx
packages/core/docs/content/locales/ja-JP/cli-adapters.mdx
- Access to secretsline 99
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/locales/ja-JP/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">Component</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 126
optional: true, description: "[useQuery オプション](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) のすべて (`queryKey` と `queryFn` を除く。これらはフック自体が設定します)。", - Network accessline 127
description: "[useQuery オプション](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) のすべて (`queryKey` と `queryFn` を除く。これらはフック自体が設定します)。", }, - Network accessline 128
"[useQuery オプション](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) のすべて (`queryKey` と `queryFn` を除く。これらはフック自体が設定します)。", - Network accessline 138
このフックは、読み取り専用のアクションで使用することを意図しています。[useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 関数の薄いラッパーです。 - Network accessline 174
optional: true, description: "[useMutation オプション](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) のすべて (`mutationFn` を除く。これはフック自体が設定します)。", - Network accessline 175
description: "[useMutation オプション](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) のすべて (`mutationFn` を除く。これはフック自体が設定します)。", fields: [ - Network accessline 176
"[useMutation オプション](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) のすべて (`mutationFn` を除く。これはフック自体が設定します)。", - Network accessline 202
このフックは、作成、更新、送信など、データを変更するアクションで使用することを意図しています。[useMutation()](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 関数の薄いラッパーです。
packages/core/docs/content/locales/ja-JP/cloneable-saas.mdx
- Network accessline 48
https://github.com/owner/repo - Network accessline 66
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 67
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 76
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 77
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 148
掲載を希望しますか?リポジトリ URL、1 文の説明、メンテナーとライセンスの 詳細、テスト済みの CLI コマンド、任意のホスト済みデモ URL を添えて、 [コミュニティテンプレートを送信](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) - Network accessline 149
詳細、テスト済みの CLI コマンド、任意のホスト済みデモ URL を添えて、 [コミュニティテンプレートを送信](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) してください。 - Network accessline 150
[コミュニティテンプレートを送信](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) - Access to secretsline 303
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/cloudflare.mdx
- Access to secretsline 79
[Credential-related source omitted] - Network accessline 91
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 131
[Credential-related source omitted] - Network accessline 166
スケジュールされた自動化、キューに入ったイベント自動化、中断された実行のクリーンアップは、すべて 1 分に 1 回のスイープで行われます。ビルドは生成される `wrangler.json` に `* * * * *` の [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) を追加し、すでに宣言されている Cron Trigger はそのまま残します。トリガ
packages/core/docs/content/locales/ja-JP/code-agents-ui.mdx
- Network accessline 205
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 206
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 358
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 359
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard"
packages/core/docs/content/locales/ja-JP/create-with-an-agent.mdx
- Network accessline 14
- ターミナルコマンドを実行できるAIエージェント。たとえばClaude Code、Codex、Cursorです。ChatGPTやWeb版Claudeのようにチャットしかできないエージェントでも手伝えます。[コマンドを実行できないエージェント](#chat-only-agents)を参照してください。 - [Node.js](https://nodejs.org) 22.22以降。 - Network accessline 15
- [Node.js](https://nodejs.org) 22.22以降。 - Network accessline 42
ください。 6. `pnpm dev` が表示するローカル URL を読み取ってください。 `curl -I <url>` などで応答があることを確認してください。ブラウザは - Network accessline 43
6. `pnpm dev` が表示するローカル URL を読み取ってください。 `curl -I <url>` などで応答があることを確認してください。ブラウザは 開かず、ヘッドレスブラウザのツールも使わないでください。アプリは - Network accessline 44
`curl -I <url>` などで応答があることを確認してください。ブラウザは - Access to secretsline 55
[Credential-related source omitted] - Network accessline 64
- 新しいアプリのフォルダ。名前は変更していなければ`my-app`で、`actions/`、`app/`、`AGENTS.md`ファイルを含みます。 - エラーなく動作している`pnpm dev`。 - エージェントが応答することを確認したローカルURL。通常は`http://localhost:8080`です。 - Network accessline 66
- エージェントが応答することを確認したローカルURL。通常は`http://localhost:8080`です。
packages/core/docs/content/locales/ja-JP/creating-templates.mdx
packages/core/docs/content/locales/ja-JP/decisions.mdx
- Network accessline 18
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 19
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 20
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/locales/ja-JP/deno-deploy.mdx
- Network accessline 20
Deno Deploy アカウントと、自分のコンピューターにインストールした [Deno](https://docs.deno.com/runtime/getting_started/installation/) も必要です。このガイドのコマンドは自分のコンピューター上で実行してください。アプリは Node.js でビルドし、`deno deploy` コマンドが完成したビルドを Deno Deploy にアップロードします。 - Access to secretsline 79
[Credential-related source omitted] - Network accessline 91
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 131
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/deployment-environment-variables.mdx
- Network accessline 28
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | 自動サーバーレススキップの対象外となるデプロイで、オンデマンドのスキーマ DDL をスキップします。スキーマがすでにマイグレーション済みであることが前提です。[オンデマンドのスキーマ DDL](/docs/deployment#skip-ensure-tables) を参照。 - Network accessline 30
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | 企業コールバック専用の、信頼された正確なプレビューオリジンのカンマ区切り許可リスト(例: `https://0123456789abcdef01234567--content.netlify.app`)です。企業コールバックデプロイにのみ設定します。Netlify のオリジンは、変更不可なデプロイ固有の 24 桁 16 進パーマリンクを使用する必要があります。変更可能な `depl - Network accessline 39
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 40
| `BETTER_AUTH_SECRET` | 32 文字以上のランダムな文字列。セッション Cookie に署名し、`OAUTH_STATE_SECRET` とアプリローカルな暗号化値のフォールバックになります。既存の共有 vault 暗号文の読み取り候補としてのレガシー用途もあります。必須項目です。本番で欠けているとフレームワークは起動時に例外をスローします。 - Network accessline 41
| `BETTER_AUTH_URL` | このアプリの公開オリジンを上書きする任意の値(例: `https://mail.example.com`)です。未設定の場合、フレームワークはリクエスト、既知のテンプレート、ホスティングプラットフォームのコンテキストから推測します。 - Access to secretsline 42
[Credential-related source omitted] - Access to secretsline 51
[Credential-related source omitted] - Access to secretsline 61
[Credential-related source omitted] - Access to secretsline 63
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/deployment.mdx
packages/core/docs/content/locales/ja-JP/dispatch.mdx
- Network accessline 116
ディスパッチは、外部エージェントの単一の MCP コネクタにすることができます。Claude、ChatGPT、Codex、または Cursor に `https://dispatch.agent-native.com/mcp` を 1 回追加すると、アプリごとに 1 つのコネクタではなく、付与されたすべてのワークスペース アプリに 1 つの承認が到達します。完全な接続フロー、アプリ許可、OAuth、およびインライン MCP アプリ プレビューについては、[External A - Network accessline 125
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p
packages/core/docs/content/locales/ja-JP/docker.mdx
- Dangerous shell operationline 56
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 58
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 95
[Credential-related source omitted] - Access to secretsline 138
[Credential-related source omitted] - Network accessline 150
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/locales/ja-JP/embedding-sdk.mdx
- Network accessline 222
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 223
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 224
url="https://assets.agent-native.com/picker" - Network accessline 277
getMcpUrl("https://assets.agent-native.com"); - Network accessline 278
getA2AUrl("https://assets.agent-native.com"); - Network accessline 280
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 281
"https://assets.agent-native.com", - Network accessline 302
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 303
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 304
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 386
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 388
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 390
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 391
hostOrigin: "https://app.example.com", - Network accessline 394
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 395
hostOrigin: "https://app.example.com", - Network accessline 399
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 400
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 401
{ hostOrigin: "https://app.example.com" }, - Network accessline 405
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 406
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 407
{ hostOrigin: "https://app.example.com" }, - Network accessline 410
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 411
hostOrigin: "https://app.example.com", - Network accessline 466
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 467
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 468
hostOrigin: "https://app.example.com",
packages/core/docs/content/locales/ja-JP/extensions.mdx
- Network accessline 238
| `dbQuery(sql, args)` | SQL から読み取り (ユーザーに自動スコープ設定) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` | Legacy trusted SQL w - Network accessline 239
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` | シ - Network accessline 240
| `extensionFetch(url, options)` | シークレットを使用して安全なプロキシを介して外部 API にアクセスする | `extensionFetch('https://api.github.com/user')` | - Network accessline 268
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 269
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/locales/ja-JP/external-agents.mdx
- Network accessline 112
https://dispatch.agent-native.com/mcp - Network accessline 123
https://mail.agent-native.com/mcp - Network accessline 127
ホストされているすべてのアプリには、ヘルパー ページもあります `https://<app>/mcp/connect` とコピー可能な URL および - Network accessline 128
`https://<app>/mcp/connect` とコピー可能な URL および - Network accessline 222
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 266
生の `skills` CLI は、`SKILL.md` ファイルのみをインストールします。ローカル MCP クライアントはまだ `npx @agent-native/core@latest connect https://assets.agent-native.com` などのコネクタが必要です。 - Network accessline 267
`npx @agent-native/core@latest connect https://assets.agent-native.com` などのコネクタが必要です。 - Network accessline 273
デフォルトのクライアント選択は、サポートされているすべてのローカル クライアントです。 `--client codex`、`--client claude-code`、または別の特定のターゲットを追加してセットアップを絞り込みます。インライン表示に対応するホスト (ChatGPT、Web・デスクトップ・モバイルの Claude チャット、Claude Desktop の Code タブ) では、ピッカーやバリアントのグリッドがチャット内に表示されます。CLI/リンク専用ホスト - Network accessline 279
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 294
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 297
`reconnect` は、指定されたホストおよび選択されたクライアント (コネクタ名に関係なく、URL で一致) について、URL が `/mcp` で終わる MCP 構成エントリを検索し、インストールされている skills に手を触れたり、完全なインストール フローを再実行したりすることなく、認証マテリアルを更新または置換します。ベースアプリ URL (例: `https://plan.agent-native.com`) を渡します — `/mcp` サフィックスが推 - Network accessline 305
リモート OAuth URL を直接追加できない MCP クライアントの場合は、ブラウザでアプリを開いて、**接続** アフォーダンス (`https://<app>/mcp/connect` で提供) を使用します。ログインした状態で、**接続 / 認証** をクリックします。このページでは、検出されたエージェントを構成するワンクリックのディープ リンク、またはすぐに貼り付けられる `.mcp.json` ブロックのいずれかを提供します。 - Network accessline 311
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 312
[Credential-related source omitted] - Network accessline 313
"url": "https://mail.agent-native.com/mcp", - Network accessline 329
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 330
https://dispatch.agent-native.com/mcp - Network accessline 333
これは、`npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code` があなたのために書くのと同じ URL 専用のエントリです。次に、Claude コードで `/mcp` を実行し、**認証** を選択します。クライアントは、MCP サーバーの `401 WWW-Authenticate` チャレンジから認証を検出し、`/.well-known/oa - Network accessline 336
https://dispatch.agent-native.com/mcp - Network accessline 477
npx @agent-native/core@latest 接続 https://plan.agent-native.com --client codex --full-catalog - Network accessline 528
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 529
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Access to secretsline 752
[Credential-related source omitted] - Access to secretsline 768
[Credential-related source omitted] - Access to secretsline 774
[Credential-related source omitted] - Network accessline 780
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 781
[Credential-related source omitted] - Network accessline 782
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 783
[Credential-related source omitted] - Network accessline 805
`connect dev` は、同じ安定した MCP サーバー名 (`agent-native-mail`、`agent-native-calendar` など) をローカルの dev-lazy ゲートウェイに書き換えるため、ツール名は変わりません。開発エントリを書き込む前に、`~/.agent-native/connect-profiles.json` の現在の運用エントリをバックアップします。デフォルト ゲートウェイは `http://127.0.0.1:8080` です - Network accessline 832
- `npx @agent-native/core@latest connect https://dispatch.agent-native.com` を使用して独自のエージェントを Dispatch に接続します。直接アプリ URL は、1 つの独立したアプリが必要な場合にのみ使用してください。 - Access to secretsline 852
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/file-uploads.mdx
- Network accessline 131
異なる種類のバックエンドの場合にのみ、カスタムプロバイダーを登録してください。次の例では、`fetch()` を使用して汎用の HTTP API 経由でアップロードします。 - Network accessline 148
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 149
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 150
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/locales/ja-JP/getting-started.mdx
packages/core/docs/content/locales/ja-JP/harness-agents.mdx
- Network accessline 111
Agent-Native は [ACP](https://agentclientprotocol.com) (エージェント クライアント) として機能できます
packages/core/docs/content/locales/ja-JP/key-concepts.mdx
- Network accessline 325
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 326
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 327
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/locales/ja-JP/koyeb.mdx
packages/core/docs/content/locales/ja-JP/mcp-apps.mdx
- Network accessline 23
OpenAI Plugin Extensions は、モデルが呼び出す MCP ツールにアプリの起動ポイントを追加します。インライン MCP Apps が有効な場合、共有 `open_app` ツールはグローバル サイドバーと会話のサイド パネルの両方にエントリポイントを表示します。ChatGPT は `{}` を渡して呼び出し、Core は現在のアプリのホーム画面をインライン MCP App として開きます。共有 MCP サーバーを使うアプリでは個別登録は不要です。[Ope - Network accessline 260
MCP Apps をインライン表示するホストには、Web・デスクトップ・モバイルの Claude チャットと Claude Desktop の [Code](https://github.com/anthropics/claude-code/issues/103809) タブが含まれます。その他の対応ホスト: ChatGPT; VS Code GitHub Copilot (1.109+); Microsoft 365 Copilot; Cursor (2.6+); Goos - Network accessline 262
Claude Code のターミナル、Cowork、claude.ai の Web 版 Code、Codex CLI では、代わりにディープリンクを使用してください。[Cowork はまだ MCP Apps への対応を宣言しておらず、同じコネクタで使用すると、コネクタを再接続するまで Chat のインラインアプリが表示されなくなることがあります。](https://github.com/anthropics/claude-code/issues/103924) ホストが if - Access to secretsline 268
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/mcp-clients.mdx
- Network accessline 149
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 150
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 179
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 180
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 181
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 227
[Credential-related source omitted] - Network accessline 274
Atlassian プリセットは、Atlassian Rovo MCP の OAuth 2.1 エンドポイント `https://mcp.atlassian.com/v1/mcp/authv2` を使用し、Jira と Confluence の - Network accessline 275
`https://mcp.atlassian.com/v1/mcp/authv2` を使用し、Jira と Confluence の - Network accessline 284
**Write**、**Search** 権限を有効にする必要がある場合があります。 ドメインには Agent-Native アプリが実行されるホストを指定してください。 実際のオリジンでない限り、別のデプロイに `https://builder.io/**` を - Network accessline 285
ドメインには Agent-Native アプリが実行されるホストを指定してください。 実際のオリジンでない限り、別のデプロイに `https://builder.io/**` を コピーしないでください。Atlassian の [Rovo MCP セットアップガイド](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 286
実際のオリジンでない限り、別のデプロイに `https://builder.io/**` を - Network accessline 287
コピーしないでください。Atlassian の [Rovo MCP セットアップガイド](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 288
と [管理者向け制御](https://support.atlassian.com/security-and-access-policies/docs/control-atlassian-rovo-mcp-server-settings/)も参照してください。 - Network accessline 300
1. **サーバー名** — 独自の参照用の短いラベル (例: "zapier"、"slack-tools")。 2. **URL** — MCP サーバープロバイダーから提供された HTTPS エンドポイント (例: `https://mcp.zapier.com/s/abc123/mcp`)。これは通常、プロバイダーのダッシュボードまたは統合ドキュメントに記載されています。 - Network accessline 301
2. **URL** — MCP サーバープロバイダーから提供された HTTPS エンドポイント (例: `https://mcp.zapier.com/s/abc123/mcp`)。これは通常、プロバイダーのダッシュボードまたは統合ドキュメントに記載されています。 - Network accessline 316
[Credential-related source omitted] - Network accessline 408
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 409
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 410
"url": "https://example.com/mcp", - Network accessline 448
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com
packages/core/docs/content/locales/ja-JP/mcp-protocol.mdx
- Access to secretsline 31
[Credential-related source omitted] - Network accessline 108
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 117
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 118
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 119
"url": "https://mail.example.com/mcp", - Network accessline 138
POST https://your-app.example.com/mcp - Access to secretsline 143
[Credential-related source omitted] - Access to secretsline 256
[Credential-related source omitted] - Network accessline 264
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 270
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 271
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 386
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 395
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 396
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 397
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/locales/ja-JP/messaging.mdx
- Network accessline 81
1. **[api.slack.com/apps](https://api.slack.com/apps)** に移動し、**新しいアプリの作成** → **最初から** をクリックします。名前を付けて (例: 「エージェント」)、ワークスペースを選択します。 - Network accessline 97
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 120
https://your-clips.example.com/api/slack/unfurl - Access to secretsline 134
[Credential-related source omitted] - Network accessline 166
OAuth Redirect URL を `https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`、 - Network accessline 167
`https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`、 - Network accessline 168
Event Subscriptions Request URL を `https://your-app.example.com/_agent-native/integrations/slack/webhook`、 Interactivity Request URL を - Network accessline 169
`https://your-app.example.com/_agent-native/integrations/slack/webhook`、 - Network accessline 170
Interactivity Request URL を `https://your-app.example.com/_agent-native/integrations/slack/interactions` に設定します。 `app_home_opened`、`app_context_changed`、`app_mention`、`message.im` を購読します。 - Network accessline 171
`https://your-app.example.com/_agent-native/integrations/slack/interactions` に設定します。 - Access to secretsline 173
[Credential-related source omitted] - Network accessline 262
[Credential-related source omitted] - Access to secretsline 262
[Credential-related source omitted] - Access to secretsline 267
[Credential-related source omitted] - Access to secretsline 364
[Credential-related source omitted] - Network accessline 378
1. テレグラムを開いて **[@BotFather](https://t.me/BotFather)** とメッセージを送信します。 - Network accessline 385
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 404
1. **[resend.com](https://resend.com)** でサインアップしてください。無料利用枠で始めるには十分です。 - Network accessline 411
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 418
[Credential-related source omitted] - Network accessline 425
1. **[sendgrid.com](https://sendgrid.com)** でサインアップしてください。 - Network accessline 433
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 438
[Credential-related source omitted] - Network accessline 462
1. **[Meta Developer Portal](https://developers.facebook.com/)** に移動し、**アプリの作成** をクリックして、**ビジネス** タイプを選択します。 - Access to secretsline 469
[Credential-related source omitted] - Network accessline 475
コールバック URL: https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Access to secretsline 686
[Credential-related source omitted] - Access to secretsline 687
[Credential-related source omitted] - Access to secretsline 797
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/multi-app-workspace.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 100
[Credential-related source omitted] - Network accessline 245
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 246
"urls": { "design": "https://design.example.test" } - Network accessline 247
"design": "https://design.example.test" - Access to secretsline 276
[Credential-related source omitted] - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Network accessline 325
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 326
# https://your-agents.com/mail/* → apps/mail - Network accessline 327
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 328
# https://your-agents.com/forms/* → apps/forms
packages/core/docs/content/locales/ja-JP/native-chat-ui.mdx
- Network accessline 448
旧 API と新 API の対応表と移行手順は、[AgentKit チャット移行ガイド](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md)を参照してください。
packages/core/docs/content/locales/ja-JP/netlify.mdx
packages/core/docs/content/locales/ja-JP/node-js.mdx
packages/core/docs/content/locales/ja-JP/notifications.mdx
- Network accessline 192
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 193
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 194
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 211
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 212
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 213
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/locales/ja-JP/observability.mdx
- Network accessline 509
リポジトリの [`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel) には、シグナルごとのエンドポイントとエクスポーターの検証を備えた、このプラグインのより完全な版があります。npm には公開されていないため、インストールするのではなく上記のプラグインから始めてください。
packages/core/docs/content/locales/ja-JP/onboarding.mdx
- Access to secretsline 187
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/plan-plugin.mdx
- Network accessline 20
- **3 つのスキル** — `/visual-plan` (正規のエントリーポイント)、`/visual-recap`、`/visualize-repo`。 - **Plan MCP コネクタ** — `https://plan.agent-native.com` にあるホスト型アプリに対して登録されます (MCP エンドポイント `https://plan.agent-native.com/mcp`、サーバー名 `plan`)。 - Network accessline 21
- **Plan MCP コネクタ** — `https://plan.agent-native.com` にあるホスト型アプリに対して登録されます (MCP エンドポイント `https://plan.agent-native.com/mcp`、サーバー名 `plan`)。 - Network accessline 97
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 121
- **Chromium ベースのブラウザを使う。** Safari は、ホスト型の HTTPS Plan ページが `http://127.0.0.1` の localhost ブリッジを読み取ることをブロックする (混在コンテンツ/ - Network accessline 122
`http://127.0.0.1` の localhost ブリッジを読み取ることをブロックする (混在コンテンツ/ - Network accessline 219
- `--client codex|claude-code|claude-code-cli|cowork|all` — MCP 設定をどのローカルエージェント向けに書き出すか (デフォルトは `all`)。 - `--no-connect` — 認証せずにコネクタだけを登録します。後で `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` を実行するか、より絞った - Network accessline 220
- `--no-connect` — 認証せずにコネクタだけを登録します。後で `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` を実行するか、より絞った `--client` を選んでください。 - Network accessline 272
インストール後は、スキルと MCP ツールがセッションに読み込まれるよう、**新しい Codex スレッドを開始してください**。このプラグインは URL のみのコネクタ (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`) を同梱しており、`codex mcp login plan` で OAuth フロー - Network accessline 273
スレッドを開始してください**。このプラグインは URL のみのコネクタ (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`) を同梱しており、`codex mcp login plan` で OAuth フロー が実行されます。インストールと認証を 1 つのコマンドで済ませたい場合や、ローカルファイル - Network accessline 274
`https://plan.agent-native.com/mcp`) を同梱しており、`codex mcp login plan` で OAuth フロー - Network accessline 281
**古いインストールをお使いの場合:** 設定に同じ URL を指す `agent-native-plans` というエントリがまだ残っている場合、Codex では `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` を、それ以外では対象の `--client` を指定した同じコマンドを実行すると、正規の `plan` という名前に統合されます。
packages/core/docs/content/locales/ja-JP/pr-visual-recap.mdx
- Access to secretsline 198
[Credential-related source omitted] - Access to secretsline 199
[Credential-related source omitted] - Access to secretsline 200
[Credential-related source omitted] - Network accessline 217
DeepSeek を使うには、`VISUAL_RECAP_AGENT=openai-compatible`、 `VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`、`VISUAL_RECAP_MODEL=deepseek-chat`、 - Network accessline 218
`VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`、`VISUAL_RECAP_MODEL=deepseek-chat`、 - Access to secretsline 219
[Credential-related source omitted] - Access to secretsline 238
[Credential-related source omitted] - Access to secretsline 254
[Credential-related source omitted] - Network accessline 265
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 280
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 292
[Credential-related source omitted] - Access to secretsline 293
[Credential-related source omitted] - Network accessline 299
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | 変数。コピー版のみ(再利用可能な呼び出し側は無視します)。`false` に設定すると、プッシュごとのリキャップを停止します。以後は PR のオープン、再オープン、レビュー準備完了、マージ時と、`recap` ラベルが付いたときにだけリキャップします。必須ラベルを付けただけでは開始しません - Network accessline 301
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | シークレット。Plans アプリを別のオリジンでセルフホストしている場合のみ必要です。 - Network accessline 342
インストールするには、[BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) からファイルを、既存の `pr-visual-recap.yml` と - Access to secretsline 344
[Credential-related source omitted] - Instruction override signalline 439
- 直接的なコマンドやロールの指示のように読める行 (「Ignore previous instructions...」、「You are now...」、「Write the token to...」など)。 - Access to secretsline 555
[Credential-related source omitted] - Access to secretsline 620
[Credential-related source omitted] - Access to secretsline 621
[Credential-related source omitted] - Access to secretsline 622
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/privacy-and-data-rights.mdx
- Network accessline 59
- [GDPR 本文](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [欧州委員会: 個人からのリクエストへの対応](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 60
- [欧州委員会: 個人からのリクエストへの対応](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [California Attorney General: CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [CCPA 規則](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/locales/ja-JP/real-time-collaboration.mdx
- Possible obfuscationline 380
サーバー側のエージェント プレゼンス ヘルパーは、下位<U+200B><U+200B>レベルのコラボレーション パッケージに残ります。 - Network accessline 680
`GET /_agent-native/health` が結果を返します。解決処理自体が登録を実行するため、 curl 1 回で有効化と確認の両方ができます。 - Network accessline 681
curl 1 回で有効化と確認の両方ができます。 - Network accessline 684
curl -s https://your-app.example.com/_agent-native/health | jq .realtime
packages/core/docs/content/locales/ja-JP/recurring-jobs.mdx
packages/core/docs/content/locales/ja-JP/render.mdx
packages/core/docs/content/locales/ja-JP/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- `innerHTML`、`eval()`、または `document.write()` は決して使用しないでください - Access to secretsline 395
[Credential-related source omitted] - Access to secretsline 411
[Credential-related source omitted] - Access to secretsline 417
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/server-database.mdx
- Network accessline 8
すべてのエージェント ネイティブ アプリは、その状態を PostgreSQL に保存します。UI とエージェントはどちらも、同じ [Drizzle ORM](https://orm.drizzle.team) クライアントを使って、同じ actions を通じて同じテーブルを読み書きします。ライブ同期は SSE とポーリングのフォールバックを使うため、手動で更新しなくても変更がもう一方の画面に表示されます。 - Network accessline 81
[PGlite](https://pglite.dev/) はプロセス内で動作する WASM 版の Postgres データベースです。PGlite デフォルトと同じくらい簡単なセットアップを保ちながら、デプロイ前に Postgres 固有のスキーマ問題を見つけられるようにします。PGlite と同様、これはローカル専用のストレージです。本番環境や共有環境では使用しないでください。
packages/core/docs/content/locales/ja-JP/server-middleware.mdx
- Network accessline 106
条件分岐、非同期処理、レスポンスの傍受といった、さらなるミドルウェアのパターンについては、[Nitro ミドルウェアのドキュメント](https://nitro.build/guide/routing#middleware) と [H3 のドキュメント](https://h3.unjs.io) を参照してください。 - Network accessline 149
- [**ルート**](/docs/server-routes): このミドルウェアによって保護されるカスタム ルート - [**セキュリティ**](/docs/security): データのスコープ設定、アクセス ガード、完全な認証モデル - [**Nitro ミドルウェアのドキュメント**](https://nitro.build/guide/routing#middleware): ミドルウェア API とパターンの全体 - Network accessline 151
- [**Nitro ミドルウェアのドキュメント**](https://nitro.build/guide/routing#middleware): ミドルウェア API とパターンの全体
packages/core/docs/content/locales/ja-JP/server-overview.mdx
- Network accessline 206
[Nitro](https://nitro.build) は、エージェント ネイティブの HTTP 層を支えるサーバー ツールキットです。`server/` で構成するものの大半は Nitro に関するものです。ファイルベースのルート、起動プラグイン、ミドルウェアです。 - Network accessline 214
基盤となるリクエスト/レスポンス API は [H3](https://h3.unjs.io) です。カスタム ルートで `defineEventHandler`、`createError`、`getQuery` を見かけたら、それは H3 です。フレームワークが直接カバーしていない事柄については、[Nitro のドキュメント](https://nitro.build/guide) と [H3 のドキュメント](https://h3.unjs.io) がリファレンスになります
packages/core/docs/content/locales/ja-JP/server-plugins.mdx
- Network accessline 141
プラグインは、`request`、`beforeResponse`、`close` など、Nitro のあらゆるライフサイクル フックを使用できます。[Nitro プラグインのドキュメント](https://nitro.build/guide/plugins) が、フック API の全体とプラグインの契約を詳しく解説しています。 - Network accessline 162
- [**ルート**](/docs/server-routes): カスタム ファイル ルートと使いどころ - [**エージェント指示の書き方**](/docs/writing-agent-instructions): agent-chat プラグインで `systemPrompt` を作成する方法 - [**Nitro プラグインのドキュメント**](https://nitro.build/guide/plugins): プラグイン API とライフサイクル フックの全体 - Network accessline 164
- [**Nitro プラグインのドキュメント**](https://nitro.build/guide/plugins): プラグイン API とライフサイクル フックの全体
packages/core/docs/content/locales/ja-JP/server-routes.mdx
- Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/skills-guide.mdx
- Network accessline 315
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 316
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 317
"url": "https://assets.agent-native.com", - Network accessline 318
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/locales/ja-JP/template-analytics.mdx
packages/core/docs/content/locales/ja-JP/template-assets.mdx
- Network accessline 117
ライブデモ: [assets.agent-native.com](https://assets.agent-native.com)。 - Network accessline 466
<EmbeddedApp url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 467
url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 503
アセット アプリ スキルのアプリ ID は `assets` で、ホストされているのは MCP URL です `https://assets.agent-native.com/mcp`. - Network accessline 504
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/locales/ja-JP/template-brain.mdx
- Network accessline 157
ライブデモ: [brain.agent-native.com](https://brain.agent-native.com)。
packages/core/docs/content/locales/ja-JP/template-calendar-developers.mdx
- Network accessline 30
これには [Node.js](https://nodejs.org) 22.22 以降と [pnpm](https://pnpm.io) が - Network accessline 41
アプリは `http://localhost:8082` で開きます。自動的に開かない場合は、その - Network accessline 51
1. **Google Cloud プロジェクトを作成する** (すでに使えるプロジェクトがない場合): [Google Cloud Console](https://console.cloud.google.com/) で、ページ上部 - Network accessline 52
[Google Cloud Console](https://console.cloud.google.com/) で、ページ上部 - Network accessline 61
4. **認証情報を作成する**: **APIs & Services → Credentials** で「Web application」タイプの OAuth クライアント ID を作成し、 `http://localhost:8082/_agent-native/google/callback` を承認済みのリダイレクト - Network accessline 62
application」タイプの OAuth クライアント ID を作成し、 `http://localhost:8082/_agent-native/google/callback` を承認済みのリダイレクト URI として追加します (8082 はカレンダーの既定の開発用ポートです)。 - Network accessline 63
`http://localhost:8082/_agent-native/google/callback` を承認済みのリダイレクト - Access to secretsline 80
[Credential-related source omitted] - Network accessline 83
[Zoom OAuth アプリ](https://marketplace.zoom.us/) (Zoom Marketplace →
packages/core/docs/content/locales/ja-JP/template-calendar-features.mdx
packages/core/docs/content/locales/ja-JP/template-calendar-integrations.mdx
- Access to secretsline 24
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/template-calendar.mdx
packages/core/docs/content/locales/ja-JP/template-chat-developers.mdx
- Network accessline 31
これには [Node.js](https://nodejs.org) 22.22 以降が必要です。`pnpm` が `PATH` に
packages/core/docs/content/locales/ja-JP/template-chat-integrations.mdx
- Access to secretsline 22
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/template-chat.mdx
packages/core/docs/content/locales/ja-JP/template-clips-developers.mdx
- Network accessline 22
これには、先に [Node.js](https://nodejs.org) 22.22 以降と [pnpm](https://pnpm.io) をインストールしておく必要があります。まだどちらもインストールしていない場合や、LLM 接続 (Builder.io、Anthropic/OpenAI のキー、またはローカルの Ollama) を設定していない場合は、[はじめに](/docs/getting-started) を参照してください。 - Network accessline 305
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 306
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 307
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/template-clips-embed.mdx
- Network accessline 40
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 53
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 54
[Credential-related source omitted] - Network accessline 56
const response = await fetch( - Network accessline 145
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 152
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/locales/ja-JP/template-clips-features.mdx
- Network accessline 119
デスクトップ アプリは、日常のキャプチャを最もスムーズにするために作られています。ドラッグ可能なカメラ バブルを使った画面とウェブカメラの録画、録画中のライブ文字起こしまで、ブラウザのタブを開いたままにする必要はありません。[clips.agent-native.com/download](https://clips.agent-native.com/download) からダウンロードでき、macOS、Windows、Linux に対応しています。 - Network accessline 214
[Credential-related source omitted] - Network accessline 215
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/template-clips-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/template-clips.mdx
packages/core/docs/content/locales/ja-JP/template-content.mdx
- Network accessline 12
### [今すぐ試す →](https://content.agent-native.com)
packages/core/docs/content/locales/ja-JP/template-design-developers.mdx
packages/core/docs/content/locales/ja-JP/template-design-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted] - Network accessline 49
https://design.agent-native.com/mcp - Network accessline 53
してアクセスを承認します。Claude Code は同じ URL を追加した後、`/mcp` を 実行して認証します。Codex やその他の CLI/stdio クライアントは、代わりに `npx @agent-native/core@latest connect https://design.agent-native.com` を - Network accessline 54
実行して認証します。Codex やその他の CLI/stdio クライアントは、代わりに `npx @agent-native/core@latest connect https://design.agent-native.com` を 実行します。これによりブラウザ タブが開いて承認を行い、共有シークレットを - Network accessline 55
`npx @agent-native/core@latest connect https://design.agent-native.com` を - Network accessline 70
返すようになった場合、そのまま再試行し続けないでください。Claude Code で `/mcp` を実行して再接続するか、任意のターミナルから `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` を実行して - Network accessline 71
`/mcp` を実行して再接続するか、任意のターミナルから `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` を実行して ください。どちらも、何も再インストールすることなく再認証できます。 - Network accessline 72
@agent-native/core@latest reconnect https://design.agent-native.com` を実行して
packages/core/docs/content/locales/ja-JP/template-design.mdx
packages/core/docs/content/locales/ja-JP/template-dispatch.mdx
packages/core/docs/content/locales/ja-JP/template-forms-agent.mdx
- Network accessline 40
- **「これを公開してリンクをください。」** フォームが公開され、すぐに公開 URL が渡されます。 - **「匿名のカスタマーフィードバックフォームを作成して、公開リンクをください。」** フォームは最初から匿名で公開されるので、送信者情報は一切記録されません。 - **「このフォームの送信を https://example.com からだけ許可して。」** フォームはそれ以外からの送信を拒否し始めます。これは今日時点ではエージェントに頼むしかできません。ビルダーにはこ - Network accessline 42
- **「このフォームの送信を https://example.com からだけ許可して。」** フォームはそれ以外からの送信を拒否し始めます。これは今日時点ではエージェントに頼むしかできません。ビルダーにはこのためのトグルがありません。
packages/core/docs/content/locales/ja-JP/template-forms-features.mdx
- Network accessline 14
[Credential-related source omitted] - Network accessline 15
[Credential-related source omitted] - Network accessline 94
公開の送信は、通過する前に [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/) による検証を必須にできます。設定でサイトキーとシークレットキーを設定すればオンになります。Turnstile がなくても常に有効な保護が 2 つあります。ボット送信を静かに破棄するための非表示フィールドと、人間が現実的に入力できるより速く送信されたものを拒否するチェックです。 - Network accessline 110
今日時点では、ビルダー内にこれを切り替えるトグルはありません。エージェントに設定を頼んでください。例えば「このフォームの送信を https://example.com からだけ許可して」のように依頼します。許可したい各サイトについて、正確なオリジン(スキームとドメイン、例えば `https://example.com` のように、末尾のパスなし)をエージェントに伝えてください。 - Network accessline 116
[Credential-related source omitted] - Network accessline 117
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/template-forms.mdx
packages/core/docs/content/locales/ja-JP/template-mail.mdx
- Network accessline 99
ライブデモ: [mail.agent-native.com](https://mail.agent-native.com)。 - Network accessline 170
1. [Google Cloud Console](https://console.cloud.google.com/) を開いてプロジェクトを作成します。 - Network accessline 171
2. API とサービス → ライブラリで **Gmail API** を有効にします。 3. OAuth 2.0 資格情報を作成します (タイプ: Web アプリケーション)。 `http://localhost:8085/_agent-native/google/callback` を承認されたリダイレクト URI として追加します。 4. クライアント ID とクライアント シークレットを実行中のアプリの [設定] ページにコピーし、[**Google アカウントに接続 - Network accessline 172
3. OAuth 2.0 資格情報を作成します (タイプ: Web アプリケーション)。 `http://localhost:8085/_agent-native/google/callback` を承認されたリダイレクト URI として追加します。
packages/core/docs/content/locales/ja-JP/template-plan.mdx
- Network accessline 140
これにより、`/visual-plan`、`/visual-recap`、`/visualize-repo` の 3 つのスラッシュコマンドに 加えて、`https://plan.agent-native.com/mcp` にある `plan` MCP コネクタ (エージェントが対話 - Network accessline 141
加えて、`https://plan.agent-native.com/mcp` にある `plan` MCP コネクタ (エージェントが対話 - Network accessline 226
Plans のツールが `needs auth`、`Unauthorized`、`Session terminated` のいずれかを返した 場合は、そのまま再試行し続けないでください。 `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 227
場合は、そのまま再試行し続けないでください。 `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` で再認証するか、OAuth 対応のホストでは `/mcp` → **Authenticate** を再実行してください。 - Network accessline 228
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`
packages/core/docs/content/locales/ja-JP/template-slides-features.mdx
packages/core/docs/content/locales/ja-JP/template-slides-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/template-slides.mdx
packages/core/docs/content/locales/ja-JP/toolkit-capability-packages.mdx
packages/core/docs/content/locales/ja-JP/toolkit-settings.mdx
- Access to secretsline 121
[Credential-related source omitted] - Network accessline 191
この実装では、OAuth コールバックをユーザーのマシン上のループバックアドレス `127.0.0.1` に設定する必要があります。`localhost` は代わりに使えません。このセルフサービスフローは、ホスト型の `*.agent-native.com` サイトでは利用できません。パートナー承認が必要なのは、ホスト型アプリの運営者です。運営者は OpenAI の[関心フォーム](https://openai.com/form/sign-in-with-chatgpt-in - Network accessline 201
推論には公開 Responses API のみを使用してください。ChatGPT の非公開バックエンドエンドポイントと会話データは、この認可の対象外です。OpenAI はトークン共有フローをプレビューとしているため、モデルの提供状況、上限、対応 API 機能は変更される可能性があります。OpenAI の[Sign in with ChatGPT の概要](https://developers.openai.com/siwc/token-sharing-open-source)
packages/core/docs/content/locales/ja-JP/tracking.mdx
- Network accessline 83
| プロバイダ | 環境変数 | | -------------- | ---------------------------------------------------------------------------------------------- | | ポストホッグ | `POS - Network accessline 84
| -------------- | ---------------------------------------------------------------------------------------------- | | ポストホッグ | `POSTHOG_API_KEY` (必須)、`POSTHOG_HOST` (オプション、デフォルトは `https://us.i.posthog.com`) | | ミックスパネル | `MIXPANEL_TOKEN` - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Network accessline 205
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 206
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 207
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 225
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 226
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 283
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 284
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 285
fetch("https://analytics.example.com/events", { - Network accessline 343
- **バッチ処理された HTTP** — 組み込みプロバイダーはイベントをキューに入れ、10 秒ごと、または 50 個のイベントが蓄積したときのいずれか早い方でフラッシュします。これにより、データを失うことなく送信リクエストが最小限に抑えられます。 - **SDK 依存関係はありません** — すべての組み込みプロバイダーは生の `fetch()` を使用します。 PostHog SDK、Mixpanel SDK、Amplitude SDK はありません。フレームワークを軽 - Network accessline 344
- **SDK 依存関係はありません** — すべての組み込みプロバイダーは生の `fetch()` を使用します。 PostHog SDK、Mixpanel SDK、Amplitude SDK はありません。フレームワークを軽量に保ちます。
packages/core/docs/content/locales/ja-JP/turn-into-app.mdx
- Network accessline 12
- ターミナルコマンドを実行できる AI エージェント(Claude Code、Codex、Cursor など)。Web 上の Claude や ChatGPT は、代わりに作業を Builder に引き渡せます。[実行される場所](#hosts)を参照してください。 - [Node.js](https://nodejs.org)。nodejs.org のインストーラーには `npx` が含まれており、以下のインストールコマンドの実行に使われます。 - Network accessline 13
- [Node.js](https://nodejs.org)。nodejs.org のインストーラーには `npx` が含まれており、以下のインストールコマンドの実行に使われます。 - Network accessline 18
- **お使いのマシンで。** Node.js 22.22 以降(`node --version` で確認)と、`PATH` 上の [pnpm](https://pnpm.io)。`pnpm --version` が失敗する場合は、`corepack enable` を実行してください。corepack は Node.js に同梱されています。 - Network accessline 59
Builder でのホスティングには、どちらの種類のホストからでも Agent-Native Dispatch MCP コネクタ(`https://dispatch.agent-native.com/mcp`、OAuth)が必要で、その `start-workspace-app-creation` アクションを呼び出します。お使いの側では何も実行されないため、Dispatch が返したブランチがビルドされるまで、レポートは保留または未検証のままです。スキルは Dispatch - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/locales/ja-JP/updating-ui-in-production.mdx
- Network accessline 14
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment) は、エージェントが本番環境でアプリの UI を変更できる、コード対応の別個のマネージドワークフローです。Builder.io をリポジトリで使い、UI の変更を直接プロンプトします(無料プランあり)。再デプロイは不要です。
packages/core/docs/content/locales/ja-JP/upgrading-core-ui.mdx
- Access to secretsline 72
[Credential-related source omitted] - Access to secretsline 75
[Credential-related source omitted] - Network accessline 187
正規のsource URL: [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx). - Network accessline 188
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx).
packages/core/docs/content/locales/ja-JP/vercel.mdx
- Access to secretsline 77
[Credential-related source omitted] - Network accessline 89
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 113
[Credential-related source omitted] - Network accessline 164
スケジュールされた自動化、キューに入ったイベント自動化、中断された実行のクリーンアップは、すべて 1 分に 1 回のスイープで行われます。ビルドは、`/_agent-native/jobs/_process-sweep` のスイープを呼び出す [Vercel cron job](https://vercel.com/docs/cron-jobs) を追加します。ワークスペースビルドでは、アプリごとに 1 つの cron job が `/<app-id>/_agent-nati
packages/core/docs/content/locales/ja-JP/voice-input.mdx
packages/core/docs/content/locales/ja-JP/what-is-agent-native.mdx
packages/core/docs/content/locales/ja-JP/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Network accessline 267
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 268
[Credential-related source omitted] - Network accessline 269
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 270
[Credential-related source omitted] - Access to secretsline 271
[Credential-related source omitted] - Access to secretsline 272
[Credential-related source omitted] - Network accessline 284
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 285
[Credential-related source omitted] - Network accessline 286
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 287
[Credential-related source omitted] - Access to secretsline 288
[Credential-related source omitted] - Access to secretsline 289
[Credential-related source omitted] - Network accessline 338
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 340
APP_URL=https://workspace.example.com
packages/core/docs/content/locales/ja-JP/workspace-management.mdx
- Network accessline 90
変更を加える必要がある人全員が git に慣れているわけではありません。 [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management) は、内部で git ブランチにマップするビジュアル ブランチ モデルをサポートしています。これは、コンテンツとコピーの変更、レイアウト調整、 - Network accessline 140
重要なヒント: 個人ではなく、GitHub チーム (`@org/team`) を使用してください。フレームワークとディスパッチの変更には、常にプラットフォームのレビューが必要です。 glob 構文と複数所有者のパターンについては、[GitHub CODEOWNERS docs](https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing - Network accessline 160
次に、[actions/labeler](https://github.com/actions/labeler) アクションを追加します。完全なワークフロー YAML については、そのリポジトリの README を参照してください。 PR が開かれるか更新されると、ラベルは自動的に適用されます。 - Access to secretsline 233
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/a2a-protocol.mdx
- Access to secretsline 93
[Credential-related source omitted] - Possible obfuscationline 113
카드 엔드포인트는 공개되므로 프레임워크는 이를 제공하기 전에 ID가 사용자별 또는 조직별 통합을 표시하는 skills를 수정합니다. ID가 `mcp__user_<emailhash>_…` 또는 `mcp__org_<orgid>_…`로 시작하는 모든 스킬은 게시된 카드에서 삭제됩니다. 운영자가 제어하는 <U+200B><U+200B>stdio MCP 도구(`mcp.config.json`에서 로드됨) 및 템플릿 정의 skills는 계속 표시됩니다. 이렇게 하면 - Network accessline 117
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 118
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 119
"url": "https://analytics.example.com", - Network accessline 314
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 350
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 351
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 352
"https://analytics.example.com", - Access to secretsline 354
[Credential-related source omitted] - Network accessline 372
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", - Network accessline 373
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", "cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 374
"url": "https://example.services.ai.azure.com/a2a", - Network accessline 375
"cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 413
[Azure AI Foundry 호스팅 에이전트](https://learn.microsoft.com/en-us/azure/foundry/agents/overview)는 - Network accessline 419
Foundry는 [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint)를 - Network accessline 429
[Gemini Enterprise는 외부 A2A 에이전트를 등록할 수 있으며](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent), - Network accessline 432
`POST {url}/v1/message:send`, 스트리밍은 `POST {url}/v1/message:stream`입니다. 요청은 Google OAuth 또는 ADC bearer 토큰과 Google Cloud IAM 권한을 사용합니다. 엔드포인트 및 권한 요구 사항은 [레지스트리 A2A 엔드포인트로 에이전트 호출](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-a - Network accessline 433
요청은 Google OAuth 또는 ADC bearer 토큰과 Google Cloud IAM 권한을 사용합니다. 엔드포인트 및 권한 요구 사항은 [레지스트리 A2A 엔드포인트로 에이전트 호출](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a)을 참조하세요. - Network accessline 434
사용합니다. 엔드포인트 및 권한 요구 사항은 [레지스트리 A2A 엔드포인트로 에이전트 호출](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a)을 - Network accessline 451
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 452
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 453
"url": "https://api.anthropic.com", - Access to secretsline 458
[Credential-related source omitted] - Access to secretsline 471
[Credential-related source omitted] - Network accessline 517
# 배포된 앱을 URL로 직접 호출하고, 환경 변수에서 전달자 토큰을 읽습니다 agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 518
[Credential-related source omitted] - Access to secretsline 518
[Credential-related source omitted] - Access to secretsline 519
[Credential-related source omitted] - Access to secretsline 591
[Credential-related source omitted] - Access to secretsline 595
[Credential-related source omitted] - Network accessline 653
- 공유 `A2A_SECRET`는 빌드 시 모든 앱 환경에 마운트됩니다. - 교차 앱 호출은 동일한 출처(`https://workspace.example.com/apps/analytics`가 `https://workspace.example.com/apps/mail`를 호출함)이므로 DNS, CORS 또는 쌍별 JWT 설정이 없습니다. - Network accessline 654
- 교차 앱 호출은 동일한 출처(`https://workspace.example.com/apps/analytics`가 `https://workspace.example.com/apps/mail`를 호출함)이므로 DNS, CORS 또는 쌍별 JWT 설정이 없습니다. - Network accessline 662
**응답보다 오래 지속되는 Fire-and-forget `Promise`에 의존하지 마십시오.** 서버리스 기능(Netlify, Vercel, AWS Lambda, Cloud Run)은 응답 본문이 플러시되는 순간을 정지합니다. 때로는 기다리지 않은 `fetch(...)`의 TCP 핸드셰이크가 완료되기도 전에 말입니다. Node에서 로컬로 작동하는 패턴은 프로덕션에서 자동으로 작업을 중단합니다. - Network accessline 700
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 701
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 702
"https://analytics.example.com", - Access to secretsline 704
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/actions-defining.mdx
- Network accessline 226
`schema`는 [Standard Schema](https://standardschema.dev) 호환 라이브러리라면 무엇이든 받아들입니다.
packages/core/docs/content/locales/ko-KR/add-ui-to-your-agent.mdx
packages/core/docs/content/locales/ko-KR/agent-mentions.mdx
packages/core/docs/content/locales/ko-KR/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/agent-web-surfaces.mdx
- Possible obfuscationline 78
기존 작업 공간 앱 구성 아래에 `agentWeb`를 추가합니다(앱의 `agent-native` 키 아래 `package.json` 또는 이에 상응하는 `workspace.agentWeb`, `agentWeb` 또는 `root.agentWeb`). 공개 경로 목록은 여전히 <U+200B><U+200B>앱의 경로 액세스 설정에서 파생됩니다. `agentWeb`는 공개 화면이 상담원에게 표시되는 방식을 제어합니다. - Network accessline 171
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 172
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 173
siteUrl: "https://example.com", - Network accessline 198
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/locales/ko-KR/audit-log.mdx
- Network accessline 110
**`export-audit-ocsf`**는 조직의 트레일을 [OCSF](https://schema.ocsf.io) 1.9.0 API Activity 이벤트(`class_uid` 6003)로 오래된 순서대로 반환하며, SIEM이 주기적으로 가져가는 용도입니다. 소유자와 관리자만 사용할 수 있고, 조직 감사 로그와 동일한 `org` 및 `admins` 이벤트를 읽으며 누구의 비공개 행도 읽지 않습니다.
packages/core/docs/content/locales/ko-KR/authentication.mdx
- Network accessline 8
에이전트 기반 앱은 계정 우선 설계로 인증을 위해 [Better Auth](https://better-auth.com)를 사용합니다. 사용자는 첫 방문 시 계정을 만들고 첫날부터 실제 신원을 얻습니다. - Access to secretsline 194
[Credential-related source omitted] - Access to secretsline 199
[Credential-related source omitted] - Access to secretsline 203
[Credential-related source omitted] - Access to secretsline 208
[Credential-related source omitted] - Access to secretsline 210
[Credential-related source omitted] - Access to secretsline 232
[Credential-related source omitted] - Access to secretsline 236
[Credential-related source omitted] - Network accessline 249
https://mail.agent-native.com/mcp - Access to secretsline 474
[Credential-related source omitted] - Access to secretsline 476
[Credential-related source omitted] - Access to secretsline 478
[Credential-related source omitted] - Access to secretsline 479
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/automations.mdx
packages/core/docs/content/locales/ko-KR/aws-amplify.mdx
- Access to secretsline 87
[Credential-related source omitted] - Network accessline 99
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 154
[Credential-related source omitted] - Network accessline 168
[Amplify 콘솔](https://console.aws.amazon.com/amplify/)에서 앱을 만듭니다. - Network accessline 196
첫 빌드는 몇 분 정도 걸리며, Amplify 콘솔에서 진행 상황을 확인할 수 있습니다. 빌드가 끝나면 `https://main.d1m7bkiki6tdw1.amplifyapp.com`과 같은 형태의 앱 주소를 열고 첫 계정을 만듭니다. - Access to secretsline 233
[Credential-related source omitted] - Access to secretsline 237
[Credential-related source omitted] - Access to secretsline 242
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/aws-lambda.mdx
- Network accessline 31
AWS 계정도 필요하며, 자신의 컴퓨터에 [AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html)를 설치하고 로그인한 뒤 기본 리전을 설정해 두어야 합니다. 이 가이드의 명령은 자신의 컴퓨터에서 실행하세요. 5단계와 6단계는 AWS CLI를 사용하고, 7단계는 CloudFront 콘솔을 사용합니다. - Access to secretsline 90
[Credential-related source omitted] - Access to secretsline 116
[Credential-related source omitted] - Access to secretsline 133
[Credential-related source omitted] - Access to secretsline 176
[Credential-related source omitted] - Access to secretsline 184
[Credential-related source omitted] - Network accessline 203
첫 번째 명령은 `FunctionUrl`을 출력하며, 이 값은 `https://abc123.lambda-url.us-east-1.on.aws/`와 같은 형태입니다. 7단계에서 사용할 수 있도록 복사해 두세요. 두 `add-permission` 명령은 요청이 그 주소를 통해 함수에 도달할 수 있도록 허용합니다. - Network accessline 219
[CloudFront 콘솔](https://console.aws.amazon.com/cloudfront/v4/home)에서 CloudFront 배포를 만듭니다. - Network accessline 229
배포의 **Origins** 탭에서 오리진을 만듭니다. **Origin domain**에는 5단계의 함수 URL 호스트 이름을 `https://`와 끝의 슬래시 없이 붙여 넣으세요. **Protocol**을 **HTTPS only**로, **Response timeout**을 60초로 설정합니다. 이 오리진에는 오리진 액세스 제어를 추가하지 마세요. - Network accessline 252
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 254
[Credential-related source omitted] - Access to secretsline 256
[Credential-related source omitted] - Access to secretsline 270
[Credential-related source omitted] - Dangerous shell operationline 272
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/locales/ko-KR/azure-static-web-apps.mdx
packages/core/docs/content/locales/ko-KR/blueprint-installer.mdx
packages/core/docs/content/locales/ko-KR/cli-adapters.mdx
- Access to secretsline 99
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/locales/ko-KR/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">Component</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 126
optional: true, description: "[useQuery 옵션](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 중 `queryKey`와 `queryFn`을 제외한 모든 것입니다. 이 둘은 후크가 직접 설정합니다.", - Network accessline 127
description: "[useQuery 옵션](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 중 `queryKey`와 `queryFn`을 제외한 모든 것입니다. 이 둘은 후크가 직접 설정합니다.", }, - Network accessline 128
"[useQuery 옵션](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 중 `queryKey`와 `queryFn`을 제외한 모든 것입니다. 이 둘은 후크가 직접 설정합니다.", - Network accessline 138
이 후크는 읽기 전용 actions에 사용하기 위한 것입니다. [useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 함수를 얇게 감싼 래퍼입니다. - Network accessline 174
optional: true, description: "[useMutation 옵션](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 중 `mutationFn`을 제외한 모든 것입니다. 이는 후크가 직접 설정합니다.", - Network accessline 175
description: "[useMutation 옵션](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 중 `mutationFn`을 제외한 모든 것입니다. 이는 후크가 직접 설정합니다.", fields: [ - Network accessline 176
"[useMutation 옵션](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 중 `mutationFn`을 제외한 모든 것입니다. 이는 후크가 직접 설정합니다.", - Network accessline 202
이 후크는 생성, 수정, 전송처럼 데이터를 변경하는 actions에 사용하기 위한 것입니다. [useMutation()](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 함수를 얇게 감싼 래퍼입니다.
packages/core/docs/content/locales/ko-KR/cloneable-saas.mdx
- Network accessline 49
https://github.com/owner/repo - Network accessline 66
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 67
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 75
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 76
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 146
목록에 추가하고 싶으신가요? 리포지토리 URL, 한 문장 설명, 메인테이너와 라이선스 세부 정보, 테스트한 CLI 명령, 선택 사항인 hosted demo URL을 포함해 [커뮤니티 템플릿을 제출하세요](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml). - Network accessline 148
[커뮤니티 템플릿을 제출하세요](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml). - Access to secretsline 300
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/cloudflare.mdx
- Access to secretsline 79
[Credential-related source omitted] - Network accessline 91
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 131
[Credential-related source omitted] - Network accessline 166
예약 자동화, 대기 중인 이벤트 자동화, 중단된 실행 정리는 모두 1분에 한 번 실행되는 스윕에서 처리됩니다. 빌드는 생성된 `wrangler.json`에 `* * * * *` [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/)를 - Network accessline 167
실행되는 스윕에서 처리됩니다. 빌드는 생성된 `wrangler.json`에 `* * * * *` [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/)를 추가하며, 이미 선언한 Cron Trigger는 그대로 유지합니다. 트리거가 실행되면 Worker의 - Network accessline 168
[Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/)를
packages/core/docs/content/locales/ko-KR/code-agents-ui.mdx
- Network accessline 202
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 203
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 355
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 356
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard"
packages/core/docs/content/locales/ko-KR/context-awareness.mdx
- Possible obfuscationline 340
`navigate`는 `navigation`의 거울 이미지입니다. `navigation`가 에이전트에게 사용자의 위치를 <U+200B><U+200B>알려주는 UI인 경우 `navigate`는 UI에게 어디로 가야 하는지 알려주는 에이전트입니다. 에이전트는 애플리케이션 상태에 일회성 `navigate` 명령을 작성합니다. UI는 이를 읽고 탐색을 수행한 다음 항목을 삭제합니다.
packages/core/docs/content/locales/ko-KR/create-with-an-agent.mdx
- Network accessline 14
- Claude Code, Codex, Cursor처럼 터미널 명령을 실행할 수 있는 AI 에이전트. ChatGPT나 웹의 Claude처럼 대화만 가능한 에이전트도 도움을 줄 수 있습니다. [명령을 실행할 수 없는 에이전트](#chat-only-agents)를 참고하세요. - [Node.js](https://nodejs.org) 22.22 이상. - Network accessline 15
- [Node.js](https://nodejs.org) 22.22 이상. - Network accessline 37
5. `pnpm dev`로 앱을 시작하고 백그라운드에서 계속 실행해 줘. 6. `pnpm dev`가 출력하는 로컬 URL을 읽어 줘. 응답하는지 예를 들어 `curl -I <url>`로 확인해 줘. 브라우저를 열지 말고 - Network accessline 38
6. `pnpm dev`가 출력하는 로컬 URL을 읽어 줘. 응답하는지 예를 들어 `curl -I <url>`로 확인해 줘. 브라우저를 열지 말고 헤드리스 브라우저 도구도 쓰지 마. 앱은 내가 직접 열게. - Network accessline 39
들어 `curl -I <url>`로 확인해 줘. 브라우저를 열지 말고 - Access to secretsline 48
[Credential-related source omitted] - Network accessline 56
- `actions/`, `app/`, `AGENTS.md` 파일이 들어 있는 새 앱 폴더. 이름을 바꾸지 않았다면 `my-app`입니다. - 오류 없이 실행 중인 `pnpm dev`. - 에이전트가 응답하는지 확인한 로컬 URL. 보통 `http://localhost:8080`입니다. - Network accessline 58
- 에이전트가 응답하는지 확인한 로컬 URL. 보통 `http://localhost:8080`입니다.
packages/core/docs/content/locales/ko-KR/creating-templates.mdx
packages/core/docs/content/locales/ko-KR/decisions.mdx
- Network accessline 18
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 19
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 20
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/locales/ko-KR/deno-deploy.mdx
- Network accessline 20
Deno Deploy 계정과 컴퓨터에 설치된 [Deno](https://docs.deno.com/runtime/getting_started/installation/)도 필요합니다. 이 가이드의 명령은 자신의 컴퓨터에서 실행하세요. 앱은 Node.js로 빌드하며, 완성된 빌드는 `deno deploy` 명령이 Deno Deploy에 업로드합니다. - Access to secretsline 79
[Credential-related source omitted] - Network accessline 91
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 131
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/deployment-environment-variables.mdx
- Network accessline 30
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | 자동 서버리스 건너뛰기가 적용되지 않는 배포에서 온디맨드 스키마 DDL을 건너뜁니다. 스키마가 이미 마이그레이션되어 있어야 합니다. [On-demand schema DDL](/docs/deployment#skip-ensure-tables)을 참조하세요. - Network accessline 32
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | 기업 콜백 전용, 정확히 신뢰할 수 있는 프리뷰 오리진의 쉼표로 구분된 허용 목록입니다(예: `https://0123456789abcdef01234567--content.netlify.app`). 기업 콜백 배포에서만 설정하세요. Netlify 오리진은 변경 불가능한, 배포별 24자리 16진수 퍼머링크를 사용해야 합니다. 변경 가능한 `dep - Network accessline 45
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 46
| `BETTER_AUTH_SECRET` | 32자 이상의 무작위 문자열입니다. 세션 쿠키에 서명하며 `OAUTH_STATE_SECRET`과 앱 로컬 암호화 값의 대체값으로 사용됩니다. 기존 공유 vault 암호문에 대한 레거시 읽기 후보이기도 합니다. 필수 항목입니다. 프로덕션에서 누락되면 프레임워크가 시작 시 예외를 발생시킵니다. - Network accessline 47
| `BETTER_AUTH_URL` | 이 앱을 위한 선택적인 공개 오리진 재정의입니다(예: `https://mail.example.com`). 설정되지 않은 경우 프레임워크는 요청, 알려진 템플릿, 호스팅 플랫폼 컨텍스트로부터 이를 추론합니다. - Access to secretsline 48
[Credential-related source omitted] - Access to secretsline 57
[Credential-related source omitted] - Access to secretsline 67
[Credential-related source omitted] - Access to secretsline 69
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/deployment.mdx
packages/core/docs/content/locales/ko-KR/dispatch.mdx
- Network accessline 116
디스패치는 외부 에이전트를 위한 단일 MCP 커넥터일 수 있습니다. Claude, ChatGPT, Codex 또는 커서에 `https://dispatch.agent-native.com/mcp`를 한 번 추가하면 앱당 하나의 커넥터 대신 하나의 인증이 부여된 모든 작업 영역 앱에 도달합니다. 전체 연결 흐름, 앱 부여, OAuth 및 인라인 MCP 앱 미리보기는 [External Agents](/docs/external-agen - Network accessline 125
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p - Possible obfuscationline 260
그런 다음 자격 증명을 Vault에 추가하고 (선택적으로) **리소스** 아래에서 전역 작업 공간 리소스를 작성합니다. 액세스 모드에 따라 Vault 키를 계속 동기화하거나 부여할 수 있습니다. 모든 앱 작업 <U+200B><U+200B>영역 리소스는 자동으로 상속됩니다. 앱별 비밀 격리가 필요한 경우 개별 앱을 부여하기 전에 Vault 액세스 설정을 수동으로 전환하세요.
packages/core/docs/content/locales/ko-KR/docker.mdx
- Dangerous shell operationline 56
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 58
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 95
[Credential-related source omitted] - Access to secretsline 138
[Credential-related source omitted] - Network accessline 150
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/locales/ko-KR/embedding-sdk.mdx
- Network accessline 222
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 223
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 224
url="https://assets.agent-native.com/picker" - Network accessline 277
getMcpUrl("https://assets.agent-native.com"); - Network accessline 278
getA2AUrl("https://assets.agent-native.com"); - Network accessline 280
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 281
"https://assets.agent-native.com", - Network accessline 302
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 303
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 304
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 386
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 388
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 390
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 391
hostOrigin: "https://app.example.com", - Network accessline 394
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 395
hostOrigin: "https://app.example.com", - Network accessline 399
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 400
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 401
{ hostOrigin: "https://app.example.com" }, - Network accessline 405
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 406
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 407
{ hostOrigin: "https://app.example.com" }, - Network accessline 410
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 411
hostOrigin: "https://app.example.com", - Network accessline 462
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 463
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 464
hostOrigin: "https://app.example.com",
packages/core/docs/content/locales/ko-KR/evals.mdx
- Possible obfuscationline 191
`preprocess` 및 `analyze`는 <U+200B><U+200B>기본적으로 ID로 설정됩니다(기록원은 원시 `AgentRunOutput`를 볼 수 있음). `analyze` 단계는 LLM 지원 점수를 위해 공급자에 구애받지 않는 `judge()` 도우미와 함께 `ctx`를 받습니다.
packages/core/docs/content/locales/ko-KR/extensions.mdx
- Network accessline 238
| `dbQuery(sql, args)` | SQL에서 읽기(사용자에게 자동 범위 지정) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` | Legacy trusted SQL writes for edi - Network accessline 239
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` | 비밀이 있는 보안 프록시를 통해 - Network accessline 240
| `extensionFetch(url, options)` | 비밀이 있는 보안 프록시를 통해 외부 API에 접근 | `extensionFetch('https://api.github.com/user')` | - Network accessline 268
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 269
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/locales/ko-KR/external-agents.mdx
- Network accessline 111
https://dispatch.agent-native.com/mcp - Network accessline 122
https://mail.agent-native.com/mcp - Network accessline 126
호스팅된 모든 앱에는 다음 위치에 도우미 페이지도 있습니다. 복사 가능한 URL가 있는 `https://<app>/mcp/connect` 및 - Network accessline 127
복사 가능한 URL가 있는 `https://<app>/mcp/connect` 및 - Network accessline 225
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Possible obfuscationline 230
Claude 코드, Claude 코드 CLI, 커서, OpenCode 및 GitHub Copilot/VS 코드의 경우 `connect`는 정적 헤더 없이 표준 원격 HTTP MCP 항목을 작성합니다. 클라이언트를 다시 시작하고 메시지가 표시되면 MCP UI에서 인증합니다. Codex 및 Claude Cowork의 경우 `connect`는 <U+200B><U+200B>호환 장치 코드 흐름을 사용합니다. 즉, 앱에서 브라우저를 열고 **승인**을 한 번 - Network accessline 269
원시 `skills` CLI는 `SKILL.md` 파일만 설치합니다. 로컬 MCP 클라이언트는 여전히 `npx @agent-native/core@latest connect https://assets.agent-native.com`와 같은 커넥터가 필요합니다. - Network accessline 270
`npx @agent-native/core@latest connect https://assets.agent-native.com`와 같은 커넥터가 필요합니다. - Network accessline 276
기본 클라이언트 선택은 지원되는 모든 로컬 클라이언트입니다. 설정 범위를 좁히려면 `--client codex`, `--client claude-code` 또는 다른 특정 대상을 추가하세요. 인라인 호스트(ChatGPT, 웹·데스크톱·모바일의 Claude 채팅, Claude Desktop의 Code 탭)는 채팅에서 선택기/변형 그리드를 렌더링합니다. CLI/링크 전용 호스트(Codex CLI, Claude Code 터미널, C - Network accessline 282
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 297
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 300
`reconnect`는 지정된 호스트 및 선택한 클라이언트(커넥터 이름에 관계없이 URL와 일치)에 대해 URL가 `/mcp`로 끝나는 MCP 구성 항목을 찾은 다음 설치된 skills를 건드리거나 전체 설치 흐름을 다시 실행하지 않고 인증 자료를 새로 고치거나 교체합니다. 기본 앱 URL(예: `https://plan.agent-native.com`)를 전달하면 `/mcp` 접미사가 유추됩니다. 인증 및 도구 로드는 클라이언 - Network accessline 308
원격 OAuth URL를 직접 추가할 수 없는 MCP 클라이언트의 경우 브라우저에서 앱을 열고 **Connect** 어포던스(`https://<app>/mcp/connect`에서 제공)를 사용하세요. 로그인한 상태에서 **연결 / 승인**을 클릭하세요. 페이지에서는 감지된 에이전트를 구성하는 원클릭 딥 링크 또는 바로 붙여넣을 수 있는 `.mcp.json` 블록을 제공합니다. - Network accessline 314
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 315
[Credential-related source omitted] - Network accessline 316
"url": "https://mail.agent-native.com/mcp", - Network accessline 332
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 333
https://dispatch.agent-native.com/mcp - Network accessline 336
이는 `npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code`가 귀하를 위해 작성하는 것과 동일한 URL 전용 항목입니다. 그런 다음 Claude 코드에서 `/mcp`를 실행하고 **인증**을 선택합니다. 클라이언트는 MCP 서버의 `401 WWW-Authenticate` 챌린지에서 인증을 검색하고, `/.well- - Network accessline 339
https://dispatch.agent-native.com/mcp - Network accessline 480
npx @agent-native/core@latest 연결 https://plan.agent-native.com --client codex --full-catalog - Network accessline 533
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 534
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Possible obfuscationline 748
`GET` + `readOnly`는 <U+200B><U+200B>작업을 부작용 없이 유지하고 화면 새로 고침 변경 이벤트에서 제외됩니다. `publicAgent`는 **명시적인 선택**입니다. 공개 웹 경로는 공개 MCP/A2A 노출을 의미하지 않습니다. [Actions](/docs/actions-overview)를 참조하세요. 디자인/컨텐츠는 actions MUST 읽기 **라이브** 상태(오래된 DB 스냅샷 열이 아닌 Yjs 공동 작업 문서)를 수 - Access to secretsline 763
[Credential-related source omitted] - Access to secretsline 779
[Credential-related source omitted] - Access to secretsline 785
[Credential-related source omitted] - Network accessline 791
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 792
[Credential-related source omitted] - Network accessline 793
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 794
[Credential-related source omitted] - Network accessline 816
`connect dev`는 <U+200B><U+200B>동일한 안정적인 MCP 서버 이름(`agent-native-mail`, `agent-native-calendar` 등)을 로컬 개발 지연 게이트웨이에 다시 작성하므로 도구 이름은 변경되지 않습니다. 개발 항목을 작성하기 전에 `~/.agent-native/connect-profiles.json`의 현재 프로덕션 항목을 백업합니다. 기본 게이트웨이는 `http://127.0.0.1:8080`입니다. - Possible obfuscationline 816
`connect dev`는 <U+200B><U+200B>동일한 안정적인 MCP 서버 이름(`agent-native-mail`, `agent-native-calendar` 등)을 로컬 개발 지연 게이트웨이에 다시 작성하므로 도구 이름은 변경되지 않습니다. 개발 항목을 작성하기 전에 `~/.agent-native/connect-profiles.json`의 현재 프로덕션 항목을 백업합니다. 기본 게이트웨이는 `http://127.0.0.1:8080`입니다. - Possible obfuscationline 830
또한 전체 앱 삽입은 MCP 전달자 토큰을 브라우저에 전달하지 않습니다. MCP 호출자는 SQL에 일회성 포함 티켓을 발행합니다. iframe 실행 경로는 이를 소비하고 수명이 짧은 iframe 안전 브라우저 세션 쿠키를 설정합니다. 랜딩 URL는 <U+200B><U+200B>제3자 쿠키가 차단될 때 클라이언트가 이를 캡처하고, 주소 표시줄에서 제거하고, 동일한 출처 `fetch` 호출에 첨부할 수 있을 만큼만 임시 `__an_embed_token` - Network accessline 843
- `npx @agent-native/core@latest connect https://dispatch.agent-native.com`를 사용하여 자신의 에이전트를 Dispatch에 연결하세요. 하나의 격리된 앱을 원하는 경우에만 직접 앱 URL를 사용하세요. - Access to secretsline 863
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/file-uploads.mdx
- Network accessline 131
사용자 지정 공급자는 다른 종류의 백엔드를 사용할 때만 등록하세요. 다음 예시는 `fetch()`를 사용하여 일반 HTTP API를 통해 업로드합니다. - Network accessline 148
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 149
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 150
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/locales/ko-KR/getting-started.mdx
packages/core/docs/content/locales/ko-KR/harness-agents.mdx
- Network accessline 111
Agent-Native는 [ACP](https://agentclientprotocol.com)(에이전트 클라이언트) 역할을 할 수 있습니다.
packages/core/docs/content/locales/ko-KR/key-concepts.mdx
- Network accessline 329
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 330
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 331
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/locales/ko-KR/koyeb.mdx
packages/core/docs/content/locales/ko-KR/mcp-apps.mdx
- Network accessline 23
OpenAI Plugin Extensions는 모델이 호출하는 MCP 도구에 앱 실행 진입점을 추가합니다. 인라인 MCP Apps가 활성화되면 공유 `open_app` 도구가 전역 사이드바와 대화 사이드 패널에 진입점을 제공합니다. ChatGPT는 `{}` 인수로 도구를 호출하고 Core는 현재 앱의 홈 화면을 인라인 MCP App으로 엽니다. 공유 MCP 서버를 사용하는 앱은 별도 등록이 필요하지 않습니다. [OpenAI P - Possible obfuscationline 37
드물게 올바른 대상은 전체 앱 셸 대신 하나의 공유 React 구성 요소를 렌더링하는 집중된 앱 경로입니다. Analytics의 `/chart` 경로가 모델입니다. URL에서 컴팩트한 `SqlPanel` 페이로드를 가져와 대시보드에서 사용하는 것과 동일한 차트 구성 요소를 렌더링합니다. 이것은 여전히 <U+200B><U+200B>일반 HTML MCP 앱이 아닌 앱 삽입입니다. 일반 작업/`open_app({ path, embed: true })`를 통 - Possible obfuscationline 222
Claude 웹은 단일 프레임 이식 경로를 사용합니다. 리소스 문서는 서명된 앱 HTML를 가져와 Claude의 MCP 앱 iframe 내에서 이를 수화합니다. 왜냐하면 Claude는 <U+200B><U+200B>앱 소유 하위 iframe 또는 외부 프레임 탐색을 안정적으로 허용하지 않기 때문입니다. ChatGPT 웹은 Apps 브리지가 안정적인 `window.openai` 호스트 API 및 제한된 높이 제어를 제공하기 때문에 제어된 경로 iframe - Network accessline 261
MCP Apps를 인라인으로 렌더링하는 호스트에는 웹·데스크톱·모바일의 Claude 채팅과 Claude Desktop의 [Code](https://github.com/anthropics/claude-code/issues/103809) 탭이 포함됩니다. 그 밖의 지원 클라이언트: ChatGPT; VS Code GitHub Copilot (1.109+); Microsoft 365 Copilot; Cursor (2.6+); Goos - Network accessline 263
Claude Code 터미널, Cowork, claude.ai 웹의 Code 및 Codex CLI에서는 딥 링크 대체 수단을 사용하세요. [Cowork는 아직 MCP Apps 지원을 선언하지 않으며, 같은 커넥터에서 사용하면 커넥터를 다시 연결할 때까지 Chat의 인라인 앱이 사라질 수 있습니다.](https://github.com/anthropics/claude-code/issues/103924) 호스트가 iframe을 렌더 - Access to secretsline 269
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/mcp-clients.mdx
- Network accessline 148
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 149
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 178
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 179
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 180
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 226
[Credential-related source omitted] - Network accessline 272
Atlassian 프리셋은 Atlassian Rovo MCP의 OAuth 2.1 엔드포인트인 `https://mcp.atlassian.com/v1/mcp/authv2`를 사용하며 Jira 및 Confluence - Network accessline 273
`https://mcp.atlassian.com/v1/mcp/authv2`를 사용하며 Jira 및 Confluence - Network accessline 282
**Write**, **Search** 권한을 활성화해야 할 수 있습니다. 도메인은 Agent-Native 앱이 실행되는 호스트여야 합니다. 실제 오리진이 아니라면 다른 배포에 `https://builder.io/**`를 복사하지 마세요. Atlassian의 [Rovo MCP 설정 - Network accessline 283
Agent-Native 앱이 실행되는 호스트여야 합니다. 실제 오리진이 아니라면 다른 배포에 `https://builder.io/**`를 복사하지 마세요. Atlassian의 [Rovo MCP 설정 가이드](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/)와 - Network accessline 284
배포에 `https://builder.io/**`를 복사하지 마세요. Atlassian의 [Rovo MCP 설정 - Network accessline 285
가이드](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/)와 - Network accessline 286
[관리자 제어](https://support.atlassian.com/security-and-access-policies/docs/control-atlassian-rovo-mcp-server-settings/)를 참조하세요. - Network accessline 297
1. **서버 이름** — 참조용 짧은 라벨(예: "zapier", "slack-tools"). 2. **URL** — MCP 서버 공급자가 제공한 HTTPS 엔드포인트(예: `https://mcp.zapier.com/s/abc123/mcp`). 이는 일반적으로 제공업체의 대시보드 또는 통합 문서에서 찾을 수 있습니다. - Network accessline 298
2. **URL** — MCP 서버 공급자가 제공한 HTTPS 엔드포인트(예: `https://mcp.zapier.com/s/abc123/mcp`). 이는 일반적으로 제공업체의 대시보드 또는 통합 문서에서 찾을 수 있습니다. - Network accessline 313
[Credential-related source omitted] - Network accessline 405
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 406
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 407
"url": "https://example.com/mcp", - Network accessline 445
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com
packages/core/docs/content/locales/ko-KR/mcp-protocol.mdx
- Access to secretsline 31
[Credential-related source omitted] - Network accessline 108
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 117
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 118
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 119
"url": "https://mail.example.com/mcp", - Network accessline 138
POST https://your-app.example.com/mcp - Access to secretsline 143
[Credential-related source omitted] - Access to secretsline 262
[Credential-related source omitted] - Network accessline 270
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 276
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 277
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 392
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 401
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 402
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 403
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/locales/ko-KR/messaging.mdx
- Network accessline 81
1. **[api.slack.com/apps](https://api.slack.com/apps)**로 이동하여 **새 앱 만들기** → **처음부터**를 클릭하세요. 이름을 지정하고(예: "에이전트") 작업공간을 선택하세요. - Network accessline 97
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 120
https://your-clips.example.com/api/slack/unfurl - Access to secretsline 134
[Credential-related source omitted] - Network accessline 166
OAuth Redirect URL은 `https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 167
`https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 168
Event Subscriptions Request URL은 `https://your-app.example.com/_agent-native/integrations/slack/webhook`, Interactivity Request URL은 - Network accessline 169
`https://your-app.example.com/_agent-native/integrations/slack/webhook`, - Network accessline 170
Interactivity Request URL은 `https://your-app.example.com/_agent-native/integrations/slack/interactions`로 설정합니다. `app_home_opened`, `app_context_changed`, `app_mention`, `message.im`을 구독하세요. - Network accessline 171
`https://your-app.example.com/_agent-native/integrations/slack/interactions`로 설정합니다. - Access to secretsline 173
[Credential-related source omitted] - Network accessline 261
[Credential-related source omitted] - Access to secretsline 261
[Credential-related source omitted] - Access to secretsline 266
[Credential-related source omitted] - Access to secretsline 363
[Credential-related source omitted] - Network accessline 377
1. 텔레그램을 열고 **[@BotFather](https://t.me/BotFather)** 메시지를 보내세요. - Network accessline 384
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 403
1. **[resend.com](https://resend.com)**에 가입하세요. 무료 등급으로 시작하기에 충분합니다. - Network accessline 410
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 417
[Credential-related source omitted] - Network accessline 424
1. **[sendgrid.com](https://sendgrid.com)**에서 가입하세요. - Network accessline 433
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 440
[Credential-related source omitted] - Network accessline 464
1. **[Meta Developer Portal](https://developers.facebook.com/)**로 이동하여 **앱 만들기**를 클릭하고 **비즈니스** 유형을 선택하세요. - Access to secretsline 471
[Credential-related source omitted] - Network accessline 477
콜백 URL: https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Access to secretsline 688
[Credential-related source omitted] - Access to secretsline 689
[Credential-related source omitted] - Access to secretsline 799
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/multi-app-workspace.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 100
[Credential-related source omitted] - Network accessline 245
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 246
"urls": { "design": "https://design.example.test" } - Network accessline 247
"design": "https://design.example.test" - Access to secretsline 276
[Credential-related source omitted] - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Network accessline 325
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 326
# https://your-agents.com/mail/* → apps/mail - Network accessline 327
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 328
# https://your-agents.com/forms/* → apps/forms - Possible obfuscationline 432
자체 도메인(`mail.company.com`, `calendar.company.com`)에 있는 각 앱을 선호하시나요? 작업 공간의 모든 앱은 여전히 <U+200B><U+200B>독립적으로 배포 가능합니다. `cd apps/mail && pnpm build`는 <U+200B><U+200B>독립형 스캐폴드처럼 정확하게 작동합니다. 그런 다음 크로스 앱 A2A는 공유 `A2A_SECRET`를 사용하여 표준 JWT 서명 경로를 통과합니다. 별도로 배포된 앱 간의 도메인 간 SSO
packages/core/docs/content/locales/ko-KR/native-chat-ui.mdx
- Network accessline 447
이전 API와 새 API의 대응표 및 마이그레이션 단계는 [AgentKit 채팅 마이그레이션 가이드](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md)를 참조하세요.
packages/core/docs/content/locales/ko-KR/netlify.mdx
packages/core/docs/content/locales/ko-KR/node-js.mdx
packages/core/docs/content/locales/ko-KR/notifications.mdx
- Network accessline 192
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 193
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 194
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 211
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 212
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 213
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/locales/ko-KR/observability.mdx
- Network accessline 507
저장소의 [`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel)에는 신호별 엔드포인트와 익스포터 검증을 갖춘 이 플러그인의 더 완전한 버전이 있습니다. npm에 게시되지 않았으므로 설치하는 대신 위의 플러그인에서 시작하세요.
packages/core/docs/content/locales/ko-KR/onboarding.mdx
- Access to secretsline 186
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/plan-plugin.mdx
- Network accessline 20
- **세 개의 스킬** — `/visual-plan`(정식 진입점), `/visual-recap`, `/visualize-repo`. - **Plan MCP 커넥터** — `https://plan.agent-native.com`에서 호스팅되는 앱에 - Network accessline 21
- **Plan MCP 커넥터** — `https://plan.agent-native.com`에서 호스팅되는 앱에 - Network accessline 22
등록됩니다(MCP 엔드포인트 `https://plan.agent-native.com/mcp`, 서버 이름 - Network accessline 97
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 124
- **Chromium 계열 브라우저를 사용하세요.** Safari는 호스팅된 HTTPS Plan 페이지가 `http://127.0.0.1` localhost 브리지를 읽는 것을 차단합니다 - Network accessline 125
페이지가 `http://127.0.0.1` localhost 브리지를 읽는 것을 차단합니다 - Network accessline 230
로컬 에이전트를 지정합니다(기본값 `all`). - `--no-connect` — 인증 없이 커넥터만 등록합니다. 나중에 `npx @agent-native/core@latest connect https://plan.agent-native.com --client all`을 - Network accessline 231
- `--no-connect` — 인증 없이 커넥터만 등록합니다. 나중에 `npx @agent-native/core@latest connect https://plan.agent-native.com --client all`을 실행하거나 더 좁은 `--client`를 선택하세요. - Network accessline 232
`npx @agent-native/core@latest connect https://plan.agent-native.com --client all`을 - Network accessline 291
설치 후에는 **새 Codex 스레드를 시작**해서 스킬과 MCP 도구가 세션에 로드되게 하세요. 플러그인은 URL 전용 커넥터(`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`)를 제공하며, `codex mcp login plan`이 - Network accessline 292
하세요. 플러그인은 URL 전용 커넥터(`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`)를 제공하며, `codex mcp login plan`이 OAuth 흐름을 실행합니다. 설치와 인증을 하나의 명령으로 함께 하고 싶거나 - Network accessline 293
`https://plan.agent-native.com/mcp`)를 제공하며, `codex mcp login plan`이 - Network accessline 300
**이전 설치:** 설정에 여전히 같은 URL을 가리키는 `agent-native-plans` 항목이 있다면, Codex의 경우 `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`를, - Network accessline 301
항목이 있다면, Codex의 경우 `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`를, 다른 대상의 경우 원하는 `--client`를 붙여 같은 명령을 실행하면 정식 이름인 - Network accessline 302
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`를,
packages/core/docs/content/locales/ko-KR/pr-visual-recap.mdx
- Access to secretsline 216
[Credential-related source omitted] - Access to secretsline 217
[Credential-related source omitted] - Access to secretsline 218
[Credential-related source omitted] - Network accessline 248
DeepSeek을 사용하려면 `VISUAL_RECAP_AGENT=openai-compatible`, `VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`, - Network accessline 249
`VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`, - Access to secretsline 250
[Credential-related source omitted] - Access to secretsline 281
[Credential-related source omitted] - Access to secretsline 307
[Credential-related source omitted] - Network accessline 319
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 335
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 347
[Credential-related source omitted] - Access to secretsline 348
[Credential-related source omitted] - Network accessline 354
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | 변수, copy 변형 전용(재사용 호출자는 무시함). `false`로 설정하면 푸시할 때마다 리캡하지 않습니다. 이후에는 PR이 열리거나 다시 열리거나 리뷰 준비가 되거나 머지될 때와 `recap` 레이블이 붙었을 때만 리캡합니다. 필수 레이블만 붙여서는 시작되지 않습 - Network accessline 356
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | 시크릿. Plans 앱을 다른 origin에서 자체 호스팅할 때만 필요합니다. - Network accessline 401
설치하려면 [BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml)에서 - Network accessline 402
[BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml)에서 - Access to secretsline 405
[Credential-related source omitted] - Access to secretsline 629
[Credential-related source omitted] - Access to secretsline 698
[Credential-related source omitted] - Access to secretsline 699
[Credential-related source omitted] - Access to secretsline 700
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/privacy-and-data-rights.mdx
- Network accessline 59
- [GDPR 본문](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [유럽연합 집행위원회: 개인 요청 처리](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 60
- [유럽연합 집행위원회: 개인 요청 처리](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [California Attorney General: CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [CCPA 규정](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/locales/ko-KR/real-time-collaboration.mdx
- Possible obfuscationline 549
서버측 actions는 `agentUpdateSelection()`를 호출하여 에이전트가 작업 중인 위치를 게시합니다. 디자인 템플릿의 `edit-design` 및 `generate-design` actions는 <U+200B><U+200B>이를 자동으로 호출합니다. 다른 템플릿도 동일한 작업을 수행할 수 있습니다: - Network accessline 676
`GET /_agent-native/health`가 결과를 알려주며, 이를 해석하는 과정에서 등록이 수행되므로 curl 한 번으로 활성화와 확인이 동시에 됩니다. - Network accessline 677
수행되므로 curl 한 번으로 활성화와 확인이 동시에 됩니다. - Network accessline 680
curl -s https://your-app.example.com/_agent-native/health | jq .realtime
packages/core/docs/content/locales/ko-KR/recurring-jobs.mdx
packages/core/docs/content/locales/ko-KR/render.mdx
packages/core/docs/content/locales/ko-KR/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- 절대 `innerHTML`, `eval()` 또는 `document.write()`를 사용하지 마세요 - Access to secretsline 391
[Credential-related source omitted] - Access to secretsline 407
[Credential-related source omitted] - Access to secretsline 413
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/server-database.mdx
- Network accessline 8
모든 에이전트 기반 앱은 상태를 PostgreSQL에 저장합니다. UI와 에이전트는 동일한 [Drizzle ORM](https://orm.drizzle.team) 클라이언트를 사용해 동일한 액션을 통해 동일한 테이블을 읽고 씁니다. 실시간 동기화는 SSE를 사용하고 폴링을 fallback으로 사용하므로, 수동으로 새로 고치지 않아도 변경 사항이 다른 화면에 표시됩니다. - Network accessline 81
[PGlite](https://pglite.dev/)는 프로세스 내에서 실행되는 WASM 기반 Postgres 데이터베이스입니다. PGlite 기본값만큼 설정을 간단하게 유지하면서도 Postgres 전용 스키마 문제를 배포 전에 잡아낼 수 있게 해줍니다. PGlite와 마찬가지로 로컬 전용 저장소입니다. 프로덕션이나 공유 환경에는 사용하지 마세요.
packages/core/docs/content/locales/ko-KR/server-middleware.mdx
- Network accessline 104
조건부 로직, 비동기 작업, 응답 가로채기 등 더 많은 미들웨어 패턴은 [Nitro 미들웨어 문서](https://nitro.build/guide/routing#middleware)와 [H3 문서](https://h3.unjs.io)를 참조하세요. - Network accessline 147
- [**라우트**](/docs/server-routes): 이 미들웨어로 보호되는 커스텀 라우트 - [**Security**](/docs/security): 데이터 범위 지정, 접근 가드, 전체 인증 모델 - [**Nitro 미들웨어 문서**](https://nitro.build/guide/routing#middleware): 전체 미들웨어 API와 패턴 - Network accessline 149
- [**Nitro 미들웨어 문서**](https://nitro.build/guide/routing#middleware): 전체 미들웨어 API와 패턴
packages/core/docs/content/locales/ko-KR/server-overview.mdx
- Network accessline 204
[Nitro](https://nitro.build)는 에이전트 기반 HTTP 계층을 지원하는 서버 툴킷입니다. `server/`에서 구성하게 될 대부분은 Nitro입니다. 파일 기반 라우트, 시작 플러그인, 미들웨어가 그것입니다. - Network accessline 212
기반이 되는 요청/응답 API는 [H3](https://h3.unjs.io)입니다. 커스텀 라우트에서 `defineEventHandler`, `createError`, `getQuery`를 보게 된다면 그것이 H3입니다. 프레임워크가 직접 다루지 않는 모든 것에 대해서는 [Nitro 문서](https://nitro.build/guide)와 [H3 문서](https://h3.unjs.io)를 참고하세요.
packages/core/docs/content/locales/ko-KR/server-plugins.mdx
- Network accessline 140
플러그인은 `request`, `beforeResponse`, `close` 등 Nitro의 모든 생명주기 훅을 사용할 수 있습니다. [Nitro 플러그인 문서](https://nitro.build/guide/plugins)에서 전체 훅 API와 플러그인 계약을 자세히 다룹니다. - Network accessline 162
- [**라우트**](/docs/server-routes): 커스텀 파일 라우트와 사용 시점 - [**Writing Agent Instructions**](/docs/writing-agent-instructions): 에이전트 채팅 플러그인의 `systemPrompt`를 작성하는 법 - [**Nitro 플러그인 문서**](https://nitro.build/guide/plugins): 전체 플러그인 API와 생명주기 훅 - Network accessline 164
- [**Nitro 플러그인 문서**](https://nitro.build/guide/plugins): 전체 플러그인 API와 생명주기 훅
packages/core/docs/content/locales/ko-KR/server-routes.mdx
- Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/skills-guide.mdx
- Network accessline 315
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 316
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 317
"url": "https://assets.agent-native.com", - Network accessline 318
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/locales/ko-KR/template-analytics.mdx
packages/core/docs/content/locales/ko-KR/template-assets.mdx
- Network accessline 115
라이브 데모: [assets.agent-native.com](https://assets.agent-native.com). - Network accessline 468
<EmbeddedApp url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 469
url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 505
자산 앱 스킬에는 앱 ID가 `assets`이고 호스팅된 MCP URL가 있습니다. `https://assets.agent-native.com/mcp`. - Network accessline 506
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/locales/ko-KR/template-brain.mdx
- Network accessline 155
라이브 데모: [brain.agent-native.com](https://brain.agent-native.com).
packages/core/docs/content/locales/ko-KR/template-calendar-developers.mdx
- Network accessline 28
이를 위해서는 먼저 [Node.js](https://nodejs.org) 22.22 이상과 [pnpm](https://pnpm.io)이 - Network accessline 39
앱이 `http://localhost:8082`에서 열립니다. 자동으로 열리지 않으면 브라우저에서 해당 - Network accessline 48
1. **Google Cloud 프로젝트 만들기** (사용할 프로젝트가 아직 없다면): [Google Cloud Console](https://console.cloud.google.com/)에서 페이지 상단 - Network accessline 49
[Google Cloud Console](https://console.cloud.google.com/)에서 페이지 상단 - Network accessline 56
화면을 설정하세요(개발용으로는 "External"이면 충분합니다). 4. **자격 증명 만들기**: **APIs & Services → Credentials**에서 "Web application" 유형의 OAuth 클라이언트 ID를 만들고, `http://localhost:8082/_agent-native/google/callback`을 - Network accessline 57
4. **자격 증명 만들기**: **APIs & Services → Credentials**에서 "Web application" 유형의 OAuth 클라이언트 ID를 만들고, `http://localhost:8082/_agent-native/google/callback`을 승인된 리디렉션 URI로 추가하세요(8082는 캘린더의 기본 개발 포트입니다). - Network accessline 58
유형의 OAuth 클라이언트 ID를 만들고, `http://localhost:8082/_agent-native/google/callback`을 - Access to secretsline 73
[Credential-related source omitted] - Network accessline 76
[Zoom OAuth 앱](https://marketplace.zoom.us/)(Zoom Marketplace → Develop → Build
packages/core/docs/content/locales/ko-KR/template-calendar-features.mdx
packages/core/docs/content/locales/ko-KR/template-calendar.mdx
packages/core/docs/content/locales/ko-KR/template-chat-developers.mdx
- Network accessline 30
이를 위해서는 [Node.js](https://nodejs.org) 22.22 이상이 필요합니다. `pnpm`이
packages/core/docs/content/locales/ko-KR/template-chat.mdx
packages/core/docs/content/locales/ko-KR/template-clips-developers.mdx
- Network accessline 22
먼저 [Node.js](https://nodejs.org) 22.22 이상과 [pnpm](https://pnpm.io)이 설치되어 있어야 합니다. 아직 설치하지 않았거나 LLM 연결(Builder.io, Anthropic/OpenAI 키, 또는 로컬 Ollama)이 설정되어 있지 않다면 [시작하기](/docs/getting-started) 문서를 참고하세요. - Network accessline 305
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 306
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 307
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/template-clips-embed.mdx
- Network accessline 40
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 53
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 54
[Credential-related source omitted] - Network accessline 56
const response = await fetch( - Network accessline 145
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 152
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/locales/ko-KR/template-clips-features.mdx
- Network accessline 119
데스크톱 앱은 가장 매끄러운 일상적인 캡처를 위해 만들어졌습니다. 드래그 가능한 카메라 버블과 함께 화면과 웹캠을 녹화하고, 녹화하는 동안 실시간으로 전사하며, 브라우저 탭을 열어둘 필요도 없습니다. [clips.agent-native.com/download](https://clips.agent-native.com/download)에서 다운로드하세요. macOS, Windows, Linux에서 사용할 수 있습니다. - Network accessline 214
[Credential-related source omitted] - Network accessline 215
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/template-clips.mdx
packages/core/docs/content/locales/ko-KR/template-content.mdx
- Network accessline 17
### [지금 사용해 보기 →](https://content.agent-native.com)
packages/core/docs/content/locales/ko-KR/template-design-developers.mdx
packages/core/docs/content/locales/ko-KR/template-design-integrations.mdx
- Access to secretsline 24
[Credential-related source omitted] - Network accessline 48
https://design.agent-native.com/mcp - Network accessline 52
승인합니다. Claude Code는 같은 URL을 추가한 다음 `/mcp`를 실행해 인증합니다. Codex와 그 밖의 CLI 또는 stdio 클라이언트는 대신 `npx @agent-native/core@latest connect https://design.agent-native.com`을 - Network accessline 53
Codex와 그 밖의 CLI 또는 stdio 클라이언트는 대신 `npx @agent-native/core@latest connect https://design.agent-native.com`을 실행합니다. 그러면 브라우저 탭이 열려 승인을 받고, 공유 시크릿 없이 - Network accessline 54
@agent-native/core@latest connect https://design.agent-native.com`을 - Network accessline 68
시작하면 계속 재시도하지 마세요. Claude Code에서 `/mcp`를 실행해 다시 연결하거나, 어떤 터미널에서든 `npx -y @agent-native/core@latest reconnect https://design.agent-native.com`을 실행하세요. 둘 다 아무것도 다시 - Network accessline 69
연결하거나, 어떤 터미널에서든 `npx -y @agent-native/core@latest reconnect https://design.agent-native.com`을 실행하세요. 둘 다 아무것도 다시 설치하지 않고 재인증합니다. - Network accessline 70
https://design.agent-native.com`을 실행하세요. 둘 다 아무것도 다시
packages/core/docs/content/locales/ko-KR/template-design.mdx
packages/core/docs/content/locales/ko-KR/template-dispatch.mdx
packages/core/docs/content/locales/ko-KR/template-forms-agent.mdx
- Network accessline 40
- **"이거 게시하고 링크 줘."** 양식이 바로 게시되고 공개 URL을 즉시 받습니다. - **"익명 고객 피드백 양식을 만들고 공개 링크를 줘."** 양식이 처음부터 익명으로 게시되므로 제출자 정보가 전혀 기록되지 않습니다. - **"이 양식은 https://example.com에서만 제출되도록 해줘."** 양식이 그 외의 모든 곳에서 온 제출을 거부하기 시작합니다. 현재는 에이전트에게 요청해야만 가능합니다. 빌더에는 이 - Network accessline 42
- **"이 양식은 https://example.com에서만 제출되도록 해줘."** 양식이 그 외의 모든 곳에서 온 제출을 거부하기 시작합니다. 현재는 에이전트에게 요청해야만 가능합니다. 빌더에는 이를 위한 토글이 없습니다.
packages/core/docs/content/locales/ko-KR/template-forms-features.mdx
- Network accessline 16
[Credential-related source omitted] - Network accessline 17
[Credential-related source omitted] - Network accessline 96
공개 제출은 통과되기 전에 [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/)을 통한 인증을 요구하도록 설정할 수 있습니다. 설정에서 사이트 키와 시크릿 키를 지정해 이 기능을 켜세요. Turnstile을 사용하지 않아도 두 가지 보호 기능은 항상 켜져 있습니다. 봇의 제출을 조용히 폐기하는 숨겨진 필드와, 사람이 양식을 합리적으로 작성할 수 있는 시간 - Network accessline 112
현재 빌더에는 이를 위한 토글이 없습니다. 에이전트에게 설정을 요청하세요. 예를 들어 "이 양식은 https://example.com에서만 제출되도록 해줘"처럼 말하면 됩니다. 허용할 각 사이트마다 정확한 출처(끝에 경로가 없는 `https://example.com` 같은 스킴과 도메인)를 에이전트에게 알려주세요. - Network accessline 118
[Credential-related source omitted] - Network accessline 119
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/template-forms.mdx
packages/core/docs/content/locales/ko-KR/template-mail.mdx
- Network accessline 97
라이브 데모: [mail.agent-native.com](https://mail.agent-native.com). - Network accessline 168
1. [Google Cloud Console](https://console.cloud.google.com/)를 열고 프로젝트를 생성하세요. - Network accessline 169
2. APIs & Services → Library에서 **Gmail API**를 활성화합니다. 3. OAuth 2.0 자격 증명을 만듭니다(유형: 웹 애플리케이션). 승인된 리디렉션 URI로 `http://localhost:8085/_agent-native/google/callback`를 추가하세요. 4. 실행 중인 앱의 설정 페이지에 클라이언트 ID와 클라이언트 비밀번호를 복사한 후 **Google 계정 연결**을 클릭하세 - Network accessline 170
3. OAuth 2.0 자격 증명을 만듭니다(유형: 웹 애플리케이션). 승인된 리디렉션 URI로 `http://localhost:8085/_agent-native/google/callback`를 추가하세요.
packages/core/docs/content/locales/ko-KR/template-plan.mdx
- Network accessline 135
이 명령은 `/visual-plan`, `/visual-recap`, `/visualize-repo` 세 개의 슬래시 명령과, `https://plan.agent-native.com/mcp`에 위치한 `plan` MCP 커넥터 - Network accessline 136
명령과, `https://plan.agent-native.com/mcp`에 위치한 `plan` MCP 커넥터 - Network accessline 226
Plans 도구가 `needs auth`, `Unauthorized`, `Session terminated`를 반환한다면 계속 재시도하지 마세요. `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`로 - Network accessline 227
반환한다면 계속 재시도하지 마세요. `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`로
packages/core/docs/content/locales/ko-KR/template-slides-features.mdx
packages/core/docs/content/locales/ko-KR/template-slides.mdx
packages/core/docs/content/locales/ko-KR/toolkit-capability-packages.mdx
packages/core/docs/content/locales/ko-KR/toolkit-settings.mdx
- Access to secretsline 121
[Credential-related source omitted] - Network accessline 189
이 흐름은 사용자의 컴퓨터에 있는 루프백 주소 `127.0.0.1`에서 HTTP OAuth 콜백을 받아야 하므로 원격 호스팅 앱에서는 작동하지 않습니다. `localhost`는 `127.0.0.1`을 대신할 수 없습니다. 따라서 `*.agent-native.com`의 호스팅 사이트에서는 이 흐름을 사용할 수 없습니다. 호스팅된 앱의 운영자가 파트너 승인을 받아야 하며 [관심 양식](https://openai.com/form/s - Network accessline 199
추론에는 공개 Responses API만 사용하세요. ChatGPT의 비공개 백엔드 엔드포인트와 ChatGPT 대화 데이터는 이 승인 범위에 포함되지 않습니다. OpenAI는 토큰 공유 흐름을 미리보기로 표시하므로 모델 제공 여부, 한도, 지원 API 기능이 바뀔 수 있습니다. OpenAI의 [Sign in with ChatGPT 개요](https://developers.openai.com/siwc/token-sharing-o
packages/core/docs/content/locales/ko-KR/tracking.mdx
- Network accessline 83
| 제공자 | 환경 변수 | | ---------- | -------------------------------------------------------------------------------------- | | 포스트호그 | `POSTHOG_API_KEY`(필수), `POSTH - Network accessline 84
| ---------- | -------------------------------------------------------------------------------------- | | 포스트호그 | `POSTHOG_API_KEY`(필수), `POSTHOG_HOST`(선택사항, 기본값은 `https://us.i.posthog.com`) | | 믹스패널 | `MIXPANEL_TOKEN` - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Network accessline 205
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 206
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 207
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 225
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 226
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 283
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 284
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 285
fetch("https://analytics.example.com/events", { - Network accessline 343
- **일괄 HTTP** — 내장 공급자는 이벤트를 대기열에 추가하고 10초마다 또는 50개의 이벤트가 누적될 때(둘 중 먼저 도래하는 시점) 플러시합니다. 이렇게 하면 데이터 손실 없이 아웃바운드 요청이 최소화됩니다. - **SDK 종속성 없음** — 모든 내장 공급자는 원시 `fetch()`를 사용합니다. PostHog SDK 없음, Mixpanel SDK 없음, 진폭 SDK 없음. 프레임워크를 가볍게 유지합니다. - Network accessline 344
- **SDK 종속성 없음** — 모든 내장 공급자는 원시 `fetch()`를 사용합니다. PostHog SDK 없음, Mixpanel SDK 없음, 진폭 SDK 없음. 프레임워크를 가볍게 유지합니다.
packages/core/docs/content/locales/ko-KR/turn-into-app.mdx
- Network accessline 12
- 터미널 명령을 실행할 수 있는 AI 에이전트(예: Claude Code, Codex, Cursor). 웹의 Claude나 ChatGPT는 대신 작업을 Builder에 전달할 수 있습니다. [실행 환경](#hosts)을 참조하세요. - [Node.js](https://nodejs.org). nodejs.org의 설치 프로그램에는 아래 설치 명령을 실행하는 `npx`가 포함되어 있습니다. - Network accessline 13
- [Node.js](https://nodejs.org). nodejs.org의 설치 프로그램에는 아래 설치 명령을 실행하는 `npx`가 포함되어 있습니다. - Network accessline 18
- **내 컴퓨터에서.** Node.js 22.22 이상(`node --version`으로 확인)과 `PATH`에 있는 [pnpm](https://pnpm.io). `pnpm --version`이 실패하면 `corepack enable`을 실행하세요. corepack은 Node.js에 포함되어 있습니다. - Network accessline 59
Builder 호스팅은 어느 종류의 호스트에서든 Agent-Native Dispatch MCP 커넥터(`https://dispatch.agent-native.com/mcp`, OAuth)가 필요하며 해당 커넥터의 `start-workspace-app-creation` 액션을 호출합니다. 사용자 쪽에서는 아무것도 실행되지 않으므로, Dispatch가 반환한 브랜치가 빌드될 때까지 보고서는 보류 또는 미검증 상태로 남습니다. sk - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/locales/ko-KR/updating-ui-in-production.mdx
- Network accessline 18
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment)는 에이전트가 프로덕션에서 앱의 UI를 수정할 수 있는, 코드 작업이 가능한 별도의 관리형 워크플로입니다. 저장소와 함께 Builder.io를 사용하고(무료 등급 제공) UI 변경을 직접 프롬프트로 요청하세요.
packages/core/docs/content/locales/ko-KR/upgrading-core-ui.mdx
- Access to secretsline 70
[Credential-related source omitted] - Access to secretsline 73
[Credential-related source omitted] - Network accessline 183
정식 source URL: [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx). - Network accessline 184
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx).
packages/core/docs/content/locales/ko-KR/vercel.mdx
- Access to secretsline 77
[Credential-related source omitted] - Network accessline 89
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 113
[Credential-related source omitted] - Network accessline 164
예약 자동화, 대기 중인 이벤트 자동화, 중단된 실행 정리는 모두 1분에 한 번 실행되는 스윕에서 처리됩니다. 빌드는 `/_agent-native/jobs/_process-sweep`에서 스윕을 호출하는 [Vercel cron 작업](https://vercel.com/docs/cron-jobs)을 추가합니다. - Network accessline 165
실행되는 스윕에서 처리됩니다. 빌드는 `/_agent-native/jobs/_process-sweep`에서 스윕을 호출하는 [Vercel cron 작업](https://vercel.com/docs/cron-jobs)을 추가합니다. 작업 공간 빌드는 앱마다 하나씩 `/<app-id>/_agent-native/jobs/_process-sweep`에 - Network accessline 166
스윕을 호출하는 [Vercel cron 작업](https://vercel.com/docs/cron-jobs)을 추가합니다.
packages/core/docs/content/locales/ko-KR/voice-input.mdx
packages/core/docs/content/locales/ko-KR/what-is-agent-native.mdx
packages/core/docs/content/locales/ko-KR/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Network accessline 267
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 268
[Credential-related source omitted] - Network accessline 269
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 270
[Credential-related source omitted] - Access to secretsline 271
[Credential-related source omitted] - Access to secretsline 272
[Credential-related source omitted] - Network accessline 284
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 285
[Credential-related source omitted] - Network accessline 286
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 287
[Credential-related source omitted] - Access to secretsline 288
[Credential-related source omitted] - Access to secretsline 289
[Credential-related source omitted] - Network accessline 340
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 342
APP_URL=https://workspace.example.com
packages/core/docs/content/locales/ko-KR/workspace-management.mdx
- Network accessline 90
변경이 필요한 모든 사람이 git에 익숙하지는 않습니다. [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management)는 내부적으로 git 분기에 매핑되는 시각적 분기 모델을 지원합니다. 이는 개발 환경 없이 콘텐츠 및 복사 변경, 레이아웃 조정, 디자인 - Network accessline 140
주요 팁: 개인이 아닌 GitHub 팀(`@org/team`)을 사용하세요. 프레임워크 및 디스패치 변경에는 항상 플랫폼 검토가 필요합니다. glob 구문과 다중 소유자 패턴은 [GitHub CODEOWNERS docs](https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/ - Network accessline 160
그런 다음 [actions/labeler](https://github.com/actions/labeler) 작업을 추가합니다. 전체 워크플로 YAML는 해당 저장소의 README를 참조하세요. PR을 열거나 업데이트하면 라벨이 자동으로 적용됩니다. - Access to secretsline 233
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/a2a-protocol.mdx
- Access to secretsline 95
[Credential-related source omitted] - Network accessline 119
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 120
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 121
"url": "https://analytics.example.com", - Access to secretsline 164
[Credential-related source omitted] - Network accessline 316
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 352
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 353
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 354
"https://analytics.example.com", - Access to secretsline 356
[Credential-related source omitted] - Network accessline 374
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", - Network accessline 375
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", "cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 376
"url": "https://example.services.ai.azure.com/a2a", - Network accessline 377
"cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 415
[Os agentes hospedados do Azure AI Foundry](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 422
O Foundry pode expor um endpoint A2A por meio do [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint). - Network accessline 431
[O Gemini Enterprise pode registrar agentes A2A externos](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent), - Network accessline 434
usa `POST {url}/v1/message:send` e streaming usa `POST {url}/v1/message:stream`. As solicitações usam tokens bearer do Google OAuth ou ADC e permissões do Google Cloud IAM. Consulte [Chamar um agente usando seu endpoint A2A do registro](htt - Network accessline 435
As solicitações usam tokens bearer do Google OAuth ou ADC e permissões do Google Cloud IAM. Consulte [Chamar um agente usando seu endpoint A2A do registro](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) para os requi - Network accessline 436
Cloud IAM. Consulte [Chamar um agente usando seu endpoint A2A do registro](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a) - Network accessline 453
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 454
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 455
"url": "https://api.anthropic.com", - Access to secretsline 460
[Credential-related source omitted] - Access to secretsline 473
[Credential-related source omitted] - Network accessline 519
# Chama um aplicativo implantado diretamente por URL, lendo o token de portador a partir de uma variável de ambiente agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 520
[Credential-related source omitted] - Access to secretsline 520
[Credential-related source omitted] - Access to secretsline 521
[Credential-related source omitted] - Access to secretsline 594
[Credential-related source omitted] - Access to secretsline 598
[Credential-related source omitted] - Network accessline 656
- Um `A2A_SECRET` compartilhado é montado no ambiente de cada aplicativo no momento da compilação. - As chamadas entre aplicativos têm a mesma origem — `https://workspace.example.com/apps/analytics` chama `https://workspace.example.com/apps - Network accessline 657
- As chamadas entre aplicativos têm a mesma origem — `https://workspace.example.com/apps/analytics` chama `https://workspace.example.com/apps/mail` — portanto, não há configuração de DNS, CORS ou JWT por par. - Network accessline 665
**Nunca confie em um `Promise` do tipo "dispare e esqueça" que sobrevive à resposta.** Funções sem servidor (Netlify, Vercel, AWS Lambda, Cloud Run) congelam no momento em que o corpo da resposta é liberado - às vezes antes mesmo que o hand - Network accessline 703
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 704
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 705
"https://analytics.example.com", - Access to secretsline 707
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/actions-defining.mdx
- Network accessline 227
`schema` aceita qualquer biblioteca compatível com [Standard Schema](https://standardschema.dev):
packages/core/docs/content/locales/pt-BR/add-ui-to-your-agent.mdx
- Network accessline 27
Cole a URL do Agent Card A2A v1.0 e escolha as credenciais Entra. Para serviço a serviço use `https://ai.azure.com/.default` e o papel **Foundry Agent Consumer**. - Network accessline 45
Navegue pela [Global A2A Registry](https://www.a2a-registry.org). Verifique card, host e autenticação antes de adicionar o agente e usar um segredo.
packages/core/docs/content/locales/pt-BR/agent-mentions.mdx
packages/core/docs/content/locales/pt-BR/agent-resources.mdx
- Possible obfuscationline 95
A guia Recursos do aplicativo mostra todos os três escopos. Recursos pessoais e compartilhados/organizacionais são editáveis <U+200B><U+200B>lá. Os recursos no escopo do espaço de trabalho são somente leitura em painéis de aplicativos e editados centralm
packages/core/docs/content/locales/pt-BR/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/agent-web-surfaces.mdx
- Possible obfuscationline 22
<Diagram id="doc-block-wfes97" title={"O que uma rota pública publica"} summary={"Uma rota pública se transforma em representações amigáveis <U+200B><U+200B>aos agentes. A leitura da rota é separada da chamada de ferramentas – o acesso à ferramenta perma - Network accessline 171
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 172
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 173
siteUrl: "https://example.com", - Network accessline 198
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/locales/pt-BR/audit-log.mdx
- Network accessline 110
**`export-audit-ocsf`** retorna a trilha da organização como eventos OCSF 1.9.0 API Activity ([OCSF](https://schema.ocsf.io), `class_uid` 6003), do mais antigo ao mais recente, para um SIEM coletar periodicamente. Apenas para proprietários
packages/core/docs/content/locales/pt-BR/authentication.mdx
- Network accessline 8
Os aplicativos nativos do agente usam [Better Auth](https://better-auth.com) para autenticação com um design que prioriza a conta. Os usuários criam uma conta na primeira visita e obtêm uma identidade real desde o primeiro dia. - Access to secretsline 196
[Credential-related source omitted] - Access to secretsline 201
[Credential-related source omitted] - Access to secretsline 205
[Credential-related source omitted] - Access to secretsline 210
[Credential-related source omitted] - Access to secretsline 212
[Credential-related source omitted] - Access to secretsline 234
[Credential-related source omitted] - Access to secretsline 238
[Credential-related source omitted] - Possible obfuscationline 244
A configuração dessas variáveis <U+200B><U+200B>nunca renderiza uma página de login de token para visitantes. O login na Web permanece no Better Auth ou no seu provedor `getSession` personalizado. - Network accessline 251
https://mail.agent-native.com/mcp - Access to secretsline 476
[Credential-related source omitted] - Access to secretsline 478
[Credential-related source omitted] - Access to secretsline 480
[Credential-related source omitted] - Access to secretsline 481
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/automations.mdx
packages/core/docs/content/locales/pt-BR/aws-amplify.mdx
- Access to secretsline 113
[Credential-related source omitted] - Network accessline 130
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 192
[Credential-related source omitted] - Network accessline 206
Crie o app no [console do Amplify](https://console.aws.amazon.com/amplify/): - Network accessline 241
O primeiro build leva alguns minutos, e você pode acompanhá-lo no console do Amplify. Quando ele terminar, abra o endereço do app, que se parece com `https://main.d1m7bkiki6tdw1.amplifyapp.com`, e crie a primeira conta. - Network accessline 243
`https://main.d1m7bkiki6tdw1.amplifyapp.com`, e crie a primeira conta. - Access to secretsline 295
[Credential-related source omitted] - Access to secretsline 299
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/aws-lambda.mdx
- Network accessline 38
Você também precisa de uma conta da AWS e da [AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html) - Access to secretsline 114
[Credential-related source omitted] - Access to secretsline 149
[Credential-related source omitted] - Access to secretsline 167
[Credential-related source omitted] - Access to secretsline 215
[Credential-related source omitted] - Access to secretsline 223
[Credential-related source omitted] - Network accessline 248
O primeiro comando mostra a `FunctionUrl`, que se parece com `https://abc123.lambda-url.us-east-1.on.aws/`. Copie-a para o Passo 7. Os dois - Network accessline 249
`https://abc123.lambda-url.us-east-1.on.aws/`. Copie-a para o Passo 7. Os dois - Network accessline 268
Crie a distribuição do CloudFront no [console do CloudFront](https://console.aws.amazon.com/cloudfront/v4/home): - Network accessline 269
[console do CloudFront](https://console.aws.amazon.com/cloudfront/v4/home): - Network accessline 283
Na aba **Origins** da distribuição, crie uma origem. Em **Origin domain**, cole o nome de host da URL da função do Passo 5, sem `https://` e sem a - Network accessline 284
cole o nome de host da URL da função do Passo 5, sem `https://` e sem a - Network accessline 320
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 322
[Credential-related source omitted] - Access to secretsline 324
[Credential-related source omitted] - Access to secretsline 341
[Credential-related source omitted] - Dangerous shell operationline 343
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/locales/pt-BR/azure-static-web-apps.mdx
packages/core/docs/content/locales/pt-BR/blueprint-installer.mdx
- Network accessline 68
agent-native add <kind> <name> # print a curated blueprint agent-native add <kind> <https://docs…> # pesquisar e integrar a partir de um URL - Network accessline 69
agent-native add <kind> <https://docs…> # pesquisar e integrar a partir de um URL - Network accessline 95
agent-native add provider https://docs.example.com/api | claude
packages/core/docs/content/locales/pt-BR/cli-adapters.mdx
- Access to secretsline 99
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/locales/pt-BR/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">Componente</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 126
optional: true, description: "Tudo das [opções de useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery), exceto `queryKey` e `queryFn`, que o gancho define por conta própria.", - Network accessline 127
description: "Tudo das [opções de useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery), exceto `queryKey` e `queryFn`, que o gancho define por conta própria.", }, - Network accessline 128
"Tudo das [opções de useQuery](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery), exceto `queryKey` e `queryFn`, que o gancho define por conta própria.", - Network accessline 138
Este gancho é destinado a actions somente leitura. É um wrapper leve em torno da função [useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery). - Network accessline 174
optional: true, description: "Tudo das [opções de useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation), exceto `mutationFn`, que o gancho define por conta própria.", - Network accessline 175
description: "Tudo das [opções de useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation), exceto `mutationFn`, que o gancho define por conta própria.", fields: [ - Network accessline 176
"Tudo das [opções de useMutation](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation), exceto `mutationFn`, que o gancho define por conta própria.", - Network accessline 203
Este gancho é destinado a actions que alteram dados, como criar, atualizar ou enviar algo. É um wrapper leve em torno da função [useMutation()](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation).
packages/core/docs/content/locales/pt-BR/cloneable-saas.mdx
- Network accessline 48
https://github.com/owner/repo - Network accessline 66
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 67
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 75
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 76
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 152
Quer entrar na lista? [Envie um template da comunidade](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) - Access to secretsline 307
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/cloudflare.mdx
- Access to secretsline 98
[Credential-related source omitted] - Network accessline 115
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 165
[Credential-related source omitted] - Network accessline 209
interrompidas acontecem em uma varredura executada uma vez por minuto. O build adiciona um [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) - Network accessline 210
adiciona um [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/) `* * * * *` ao `wrangler.json` gerado e mantém os Cron Triggers que você já - Network accessline 211
[Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/)
packages/core/docs/content/locales/pt-BR/code-agents-ui.mdx
- Network accessline 205
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 206
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 358
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 359
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard"
packages/core/docs/content/locales/pt-BR/create-with-an-agent.mdx
- Network accessline 14
- Um agente de IA que consiga executar comandos de terminal, como Claude Code, Codex ou Cursor. Agentes que só conversam, como ChatGPT ou Claude na web, ainda podem ajudar. Veja [Agentes que não executam comandos](#chat-only-agents). - [Nod - Network accessline 15
- [Node.js](https://nodejs.org) 22.22 ou posterior. - Network accessline 38
5. Inicie o app com `pnpm dev` e mantenha-o rodando em segundo plano. 6. Leia a URL local que o `pnpm dev` exibe. Verifique se ela responde, por exemplo com `curl -I <url>`. Não abra um navegador e - Network accessline 39
6. Leia a URL local que o `pnpm dev` exibe. Verifique se ela responde, por exemplo com `curl -I <url>`. Não abra um navegador e não use ferramentas de navegador headless. Eu mesmo abrirei o app. - Network accessline 40
responde, por exemplo com `curl -I <url>`. Não abra um navegador e - Network accessline 58
- Uma nova pasta de app, chamada `my-app` a menos que você a tenha trocado, que contém `actions/`, `app/` e um arquivo `AGENTS.md`. - `pnpm dev` em execução sem erros. - Uma URL local, geralmente `http://localhost:8080`, que o agente verifi - Network accessline 60
- Uma URL local, geralmente `http://localhost:8080`, que o agente verificou que responde.
packages/core/docs/content/locales/pt-BR/creating-templates.mdx
- Possible obfuscationline 363
Mantenha filtros compartilháveis <U+200B><U+200B>nos parâmetros de consulta URL. A estrutura os expõe ao agente como `<current-url>` e o agente integrado pode alterá-los com `set-search-params`; `navigation` deve conter IDs semânticos e aliases, e não um - Access to secretsline 487
[Credential-related source omitted] - Network accessline 532
[Envie um modelo da comunidade](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml) quando o repositório estiver pronto. Esses modelos permanecem separados do catálogo próprio e são mantidos por seus autore
packages/core/docs/content/locales/pt-BR/decisions.mdx
- Network accessline 18
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 19
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 20
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/locales/pt-BR/deno-deploy.mdx
- Network accessline 26
Você também precisa de uma conta do Deno Deploy e do [Deno](https://docs.deno.com/runtime/getting_started/installation/) - Access to secretsline 100
[Credential-related source omitted] - Network accessline 117
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 168
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/deployment-environment-variables.mdx
- Network accessline 31
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | Pula o DDL de schema sob demanda em implantações que o pulo automático em serverless não cobre. Requer que seu schema já esteja migrado. Veja [DDL de schema sob demanda](/docs/deployment#skip- - Network accessline 33
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | Lista de permissões, separada por vírgulas e exclusiva do callback corporativo, de origens de preview exatas e confiáveis (por exemplo, `https://0123456789abcdef01234567--content.netlify.app`) - Network accessline 49
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 50
| `BETTER_AUTH_SECRET` | String aleatória de 32+ caracteres. Assina cookies de sessão e é o fallback de `OAUTH_STATE_SECRET` e de valores criptografados locais do app. Também é um candidato de leitura legado para ciph - Network accessline 51
| `BETTER_AUTH_URL` | Sobrescrita opcional da origem pública para este app (por exemplo, `https://mail.example.com`). O framework a infere a partir do contexto da requisição, do template conhecido e da plataforma d - Access to secretsline 52
[Credential-related source omitted] - Access to secretsline 53
[Credential-related source omitted] - Access to secretsline 61
[Credential-related source omitted] - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 73
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/deployment.mdx
packages/core/docs/content/locales/pt-BR/dispatch.mdx
- Network accessline 116
O Dispatch pode ser o único conector MCP para agentes externos: adicione `https://dispatch.agent-native.com/mcp` uma vez em Claude, ChatGPT, Codex ou Cursor, e uma autorização atinge cada aplicativo de espaço de trabalho concedido em vez de - Network accessline 125
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p
packages/core/docs/content/locales/pt-BR/docker.mdx
- Dangerous shell operationline 70
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 72
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 113
[Credential-related source omitted] - Access to secretsline 169
[Credential-related source omitted] - Network accessline 186
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/locales/pt-BR/drop-in-agent.mdx
- Possible obfuscationline 119
note: "`suggestions` renderiza como chips clicáveis <U+200B><U+200B>no chat vazio.",
packages/core/docs/content/locales/pt-BR/embedding-sdk.mdx
- Network accessline 224
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 225
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 226
url="https://assets.agent-native.com/picker" - Network accessline 279
getMcpUrl("https://assets.agent-native.com"); - Network accessline 280
getA2AUrl("https://assets.agent-native.com"); - Network accessline 282
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 283
"https://assets.agent-native.com", - Network accessline 304
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 305
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 306
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 388
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 390
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 392
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 393
hostOrigin: "https://app.example.com", - Network accessline 396
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 397
hostOrigin: "https://app.example.com", - Network accessline 401
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 402
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 403
{ hostOrigin: "https://app.example.com" }, - Network accessline 407
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 408
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 409
{ hostOrigin: "https://app.example.com" }, - Network accessline 412
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 413
hostOrigin: "https://app.example.com", - Network accessline 468
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 469
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 470
hostOrigin: "https://app.example.com", - Possible obfuscationline 615
Por padrão, `extensionData` usa o navegador `localStorage`, que é útil para protótipos e widgets locais. Os hosts SaaS de produção devem passar um adaptador `storage` apoiado por back-end para que os dados de extensão com escopo do usuário
packages/core/docs/content/locales/pt-BR/extensions.mdx
- Network accessline 241
| `dbQuery(sql, args)` | Ler de SQL (com escopo automático para o usuário) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` | Legacy tr - Network accessline 242
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` | Atingir A - Network accessline 243
| `extensionFetch(url, options)` | Atingir APIs externos através de um proxy seguro com segredos | `extensionFetch('https://api.github.com/user')` | - Network accessline 271
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 272
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/locales/pt-BR/external-agents.mdx
- Network accessline 111
https://dispatch.agent-native.com/mcp - Network accessline 122
https://mail.agent-native.com/mcp - Network accessline 126
Todo aplicativo hospedado também tem uma página auxiliar em `https://<app>/mcp/connect` com o URL copiável e - Network accessline 127
`https://<app>/mcp/connect` com o URL copiável e - Network accessline 232
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 256
Ao solucionar problemas de configuração MCP local, edite `Authorization`, `http_headers`, e valores de token antes de compartilhar logs. Não use curl bruto como substituto de um - Network accessline 257
e valores de token antes de compartilhar logs. Não use curl bruto como substituto de um - Network accessline 276
O `skills` CLI bruto instala apenas arquivos `SKILL.md`; clientes MCP locais ainda precisa de um conector como `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 277
precisa de um conector como `npx @agent-native/core@latest connect https://assets.agent-native.com`. - Network accessline 283
A seleção de cliente padrão é composta por todos os clientes locais suportados; adicione `--client codex`, `--client claude-code` ou outro alvo específico para restringir a configuração. Hosts inline (ChatGPT, chat do Claude na web, no desk - Network accessline 289
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 304
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 307
`reconnect` encontra qualquer entrada de configuração MCP cujo URL termina em `/mcp` para o host fornecido e o cliente selecionado (correspondente a URL, independentemente do nome do conector) e, em seguida, atualiza ou substitui o material - Network accessline 315
Para clientes MCP que não podem adicionar um OAuth URL remoto diretamente, abra o aplicativo em seu navegador e use seu recurso **Connect** (servido em `https://<app>/mcp/connect`). Enquanto estiver logado, clique em **Conectar/Autori - Network accessline 321
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 322
[Credential-related source omitted] - Network accessline 323
"url": "https://mail.agent-native.com/mcp", - Network accessline 339
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 340
https://dispatch.agent-native.com/mcp - Network accessline 343
Esta é a mesma entrada somente URL que `npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code` escreve para você. Em seguida, execute `/mcp` no código Claude e escolha **Autenticar**. O cliente descobr - Network accessline 346
https://dispatch.agent-native.com/mcp - Network accessline 488
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex --full-catalog - Network accessline 539
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 540
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Access to secretsline 773
[Credential-related source omitted] - Access to secretsline 789
[Credential-related source omitted] - Access to secretsline 795
[Credential-related source omitted] - Network accessline 801
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 802
[Credential-related source omitted] - Network accessline 803
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 804
[Credential-related source omitted] - Network accessline 826
`connect dev` reescreve os mesmos nomes de servidor MCP estáveis <U+200B><U+200B>(`agent-native-mail`, `agent-native-calendar`, etc.) para o gateway local dev-lazy, para que os nomes das ferramentas não sejam alterados. Ele faz backup das entradas de pro - Possible obfuscationline 826
`connect dev` reescreve os mesmos nomes de servidor MCP estáveis <U+200B><U+200B>(`agent-native-mail`, `agent-native-calendar`, etc.) para o gateway local dev-lazy, para que os nomes das ferramentas não sejam alterados. Ele faz backup das entradas de pro - Network accessline 853
- Conecte seu próprio agente ao Dispatch com `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; use um aplicativo direto URL somente quando quiser um aplicativo isolado. - Access to secretsline 873
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/file-uploads.mdx
- Network accessline 134
Registre um provedor personalizado apenas para um tipo diferente de back-end. Este exemplo faz upload por meio de uma API HTTP genérica usando `fetch()`: - Network accessline 135
Este exemplo faz upload por meio de uma API HTTP genérica usando `fetch()`: - Network accessline 152
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 153
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 154
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/locales/pt-BR/getting-started.mdx
packages/core/docs/content/locales/pt-BR/harness-agents.mdx
- Network accessline 111
Agent-Native pode atuar como um [ACP](https://agentclientprotocol.com) (Agente Cliente
packages/core/docs/content/locales/pt-BR/key-concepts.mdx
- Network accessline 330
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 331
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 332
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/locales/pt-BR/koyeb.mdx
packages/core/docs/content/locales/pt-BR/mcp-apps.mdx
- Network accessline 23
As OpenAI Plugin Extensions adicionam pontos de entrada às ferramentas MCP chamadas pelo modelo. Quando os MCP Apps inline estão habilitados, a ferramenta compartilhada `open_app` anuncia um ponto na barra lateral global e outro no painel l - Possible obfuscationline 222
Claude web usa um caminho de transplante de quadro único: o documento de recurso busca o aplicativo assinado HTML e o hidrata dentro do iframe do aplicativo MCP de Claude porque Claude não permite de forma confiável iframes filhos de propri - Network accessline 261
Os hosts que renderizam MCP Apps inline incluem o chat do Claude na web, no desktop e no celular, além da aba [Code](https://github.com/anthropics/claude-code/issues/103809) do Claude Desktop; ChatGPT; VS Code GitHub Copilot (1.109+); Micro - Network accessline 263
Use o link direto de fallback no terminal do Claude Code, no Cowork, no Code na web claude.ai e na CLI do Codex. [O Cowork ainda não declara suporte a MCP Apps, e usá-lo no mesmo conector pode fazer os aplicativos inline desaparecerem do Ch - Access to secretsline 269
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/mcp-clients.mdx
- Possible obfuscationline 23
<Diagram id="doc-block-1pkuuh" title={"Direção do cliente: muitas fontes, um registro de ferramenta"} summary={"Arquivos de configuração, env e UI de tempo de execução são mesclados no gerenciador MCP; suas ferramentas aparecem prefixadas e - Network accessline 148
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 149
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 178
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 179
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 180
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 226
[Credential-related source omitted] - Network accessline 274
O preset Atlassian usa o endpoint OAuth 2.1 do Atlassian Rovo MCP, `https://mcp.atlassian.com/v1/mcp/authv2`, e disponibiliza ferramentas do Jira - Network accessline 275
`https://mcp.atlassian.com/v1/mcp/authv2`, e disponibiliza ferramentas do Jira - Network accessline 284
ativar as permissões necessárias de **Read**, **Write** e **Search** em **Permissions**. O domínio deve ser o host em que o aplicativo Agent-Native é executado; não copie `https://builder.io/**` para outra implantação, a menos que - Network accessline 285
**Permissions**. O domínio deve ser o host em que o aplicativo Agent-Native é executado; não copie `https://builder.io/**` para outra implantação, a menos que essa seja a origem real dela. Consulte o [guia de configuração do Rovo MCP](https - Network accessline 286
executado; não copie `https://builder.io/**` para outra implantação, a menos que - Network accessline 287
essa seja a origem real dela. Consulte o [guia de configuração do Rovo MCP](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 288
e os [controles administrativos](https://support.atlassian.com/security-and-access-policies/docs/control-atlassian-rovo-mcp-server-settings/) da Atlassian. - Network accessline 300
1. **Nome do servidor** — um rótulo curto para sua própria referência (por exemplo, "zapier", "slack-tools"). 2. **URL** — o endpoint HTTPS que o provedor do servidor MCP lhe forneceu (por exemplo, `https://mcp.zapier.com/s/abc123/mcp`). Is - Network accessline 301
2. **URL** — o endpoint HTTPS que o provedor do servidor MCP lhe forneceu (por exemplo, `https://mcp.zapier.com/s/abc123/mcp`). Isso geralmente é encontrado no painel do provedor ou nos documentos de integração. - Network accessline 316
[Credential-related source omitted] - Network accessline 408
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 409
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 410
"url": "https://example.com/mcp", - Network accessline 448
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com - Possible obfuscationline 517
As ferramentas MCP são executadas em sua máquina com quaisquer permissões que o processo gerado tenha. Trate `mcp.config.json` como qualquer outra lista de executáveis <U+200B><U+200B>que você deseja deixar o agente dirigir. As ferramentas dos servidores
packages/core/docs/content/locales/pt-BR/mcp-protocol.mdx
- Access to secretsline 31
[Credential-related source omitted] - Network accessline 108
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 117
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 118
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 119
"url": "https://mail.example.com/mcp", - Network accessline 138
POST https://your-app.example.com/mcp - Access to secretsline 143
[Credential-related source omitted] - Access to secretsline 259
[Credential-related source omitted] - Network accessline 267
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 273
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 274
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 389
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 398
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 399
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 400
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/locales/pt-BR/messaging.mdx
- Network accessline 81
1. Vá para **[api.slack.com/apps](https://api.slack.com/apps)** e clique em **Criar novo aplicativo** → **Do zero**. Dê um nome (por exemplo, "Agente") e escolha seu espaço de trabalho. - Network accessline 97
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 120
https://your-clips.example.com/api/slack/unfurl - Access to secretsline 134
[Credential-related source omitted] - Network accessline 169
Defina OAuth Redirect URL como `https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 170
`https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 171
Event Subscriptions Request URL como `https://your-app.example.com/_agent-native/integrations/slack/webhook` e Interactivity Request URL como - Network accessline 172
`https://your-app.example.com/_agent-native/integrations/slack/webhook` e - Network accessline 173
Interactivity Request URL como `https://your-app.example.com/_agent-native/integrations/slack/interactions`. Assine `app_home_opened`, `app_context_changed`, `app_mention`, - Network accessline 174
`https://your-app.example.com/_agent-native/integrations/slack/interactions`. - Access to secretsline 177
[Credential-related source omitted] - Network accessline 270
[Credential-related source omitted] - Access to secretsline 270
[Credential-related source omitted] - Access to secretsline 275
[Credential-related source omitted] - Access to secretsline 372
[Credential-related source omitted] - Network accessline 386
1. Abra o Telegram e envie a mensagem **[@BotFather](https://t.me/BotFather)**. - Network accessline 393
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 412
1. Inscreva-se em **[resend.com](https://resend.com)**. O nível gratuito é suficiente para começar. - Network accessline 419
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 426
[Credential-related source omitted] - Network accessline 433
1. Inscreva-se em **[sendgrid.com](https://sendgrid.com)**. - Network accessline 441
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 446
[Credential-related source omitted] - Network accessline 470
1. Acesse **[Meta Developer Portal](https://developers.facebook.com/)**, clique em **Criar aplicativo** e escolha o tipo **Comercial**. - Access to secretsline 477
[Credential-related source omitted] - Network accessline 483
Retorno de chamada URL: https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Possible obfuscationline 671
Quais plataformas estão ativas dependem de quais variáveis <U+200B><U+200B>de ambiente estão definidas. O plugin registra rotas de webhook para cada uma delas em `/_agent-native/integrations/`. - Access to secretsline 694
[Credential-related source omitted] - Access to secretsline 695
[Credential-related source omitted] - Access to secretsline 805
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/multi-app-workspace.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 100
[Credential-related source omitted] - Network accessline 245
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 246
"urls": { "design": "https://design.example.test" } - Network accessline 247
"design": "https://design.example.test" - Access to secretsline 276
[Credential-related source omitted] - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Network accessline 325
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 326
# https://your-agents.com/mail/* → apps/mail - Network accessline 327
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 328
# https://your-agents.com/forms/* → apps/forms
packages/core/docs/content/locales/pt-BR/native-chat-ui.mdx
- Possible obfuscationline 186
Prefira chaves de coluna estáveis <U+200B><U+200B>e valores de linha seguros para JSON. Use `totalRows`, - Network accessline 452
Consulte o [guia de migração do chat do AgentKit](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md) para ver o mapa das APIs antigas e novas e as etapas de migração.
packages/core/docs/content/locales/pt-BR/netlify.mdx
packages/core/docs/content/locales/pt-BR/node-js.mdx
packages/core/docs/content/locales/pt-BR/notifications.mdx
- Network accessline 192
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 193
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 194
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 211
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 212
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 213
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/locales/pt-BR/observability.mdx
- Network accessline 509
O diretório [`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel) do repositório traz uma versão mais completa deste plugin, com endpoints por sinal e validação do exportador. Ele não é publicado no npm, então
packages/core/docs/content/locales/pt-BR/observational-memory.mdx
- Possible obfuscationline 135
Os padrões são conservadores. Um operador pode discar a compactação no momento da implantação com variáveis <U+200B><U+200B>de ambiente `AGENT_NATIVE_OM_*` (não é necessária a reimplantação do código do aplicativo); um valor inválido ou ausente sempre re
packages/core/docs/content/locales/pt-BR/onboarding.mdx
- Access to secretsline 187
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/plan-plugin.mdx
- Network accessline 23
- **Três habilidades** — `/visual-plan` (o ponto de entrada canônico), `/visual-recap` e `/visualize-repo`. - **O conector MCP do Plan** — registrado no aplicativo hospedado em `https://plan.agent-native.com` (endpoint MCP `https://plan.age - Network accessline 24
- **O conector MCP do Plan** — registrado no aplicativo hospedado em `https://plan.agent-native.com` (endpoint MCP `https://plan.agent-native.com/mcp`, nome do servidor `plan`). - Network accessline 101
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 130
- **Use um navegador baseado em Chromium.** O Safari bloqueia a página hospedada em HTTPS do Plan de ler a ponte localhost `http://127.0.0.1` - Network accessline 131
hospedada em HTTPS do Plan de ler a ponte localhost `http://127.0.0.1` - Network accessline 240
- `--client codex|claude-code|claude-code-cli|cowork|all` — para quais agentes locais escrever a configuração MCP (padrão `all`). - `--no-connect` — registra o conector sem autenticar; rode `npx @agent-native/core@latest connect https://pla - Network accessline 241
- `--no-connect` — registra o conector sem autenticar; rode `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` depois, ou escolha um `--client` mais específico. - Network accessline 301
Depois da instalação, **inicie uma nova thread do Codex** para que as habilidades e ferramentas MCP carreguem na sessão. O plugin distribui um conector apenas com URL (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`); - Network accessline 302
habilidades e ferramentas MCP carreguem na sessão. O plugin distribui um conector apenas com URL (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`); `codex mcp login plan` executa o fluxo OAuth. A rota do CLI universal acima - Network accessline 303
conector apenas com URL (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`); - Network accessline 312
**Instalações antigas:** se sua configuração ainda tiver uma entrada `agent-native-plans` apontando para a mesma URL, rodar `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 313
`agent-native-plans` apontando para a mesma URL, rodar `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` para o Codex, ou o mesmo comando com seu `--client` alvo, consolida isso no - Network accessline 314
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`
packages/core/docs/content/locales/pt-BR/pr-visual-recap.mdx
- Access to secretsline 232
[Credential-related source omitted] - Access to secretsline 233
[Credential-related source omitted] - Access to secretsline 234
[Credential-related source omitted] - Network accessline 267
Para o DeepSeek, defina `VISUAL_RECAP_AGENT=openai-compatible`, `VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`, - Network accessline 268
`VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`, - Access to secretsline 269
[Credential-related source omitted] - Access to secretsline 305
[Credential-related source omitted] - Access to secretsline 334
[Credential-related source omitted] - Network accessline 347
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 363
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 375
[Credential-related source omitted] - Access to secretsline 376
[Credential-related source omitted] - Network accessline 382
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | Variável, apenas para a variante de cópia (o chamador reutilizável a ignora). Defina como `false` para parar de gerar o recap a ca - Network accessline 384
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | Secret. Apenas quando o aplicativo Plans está auto-hospedado em outra origem. - Network accessline 430
Para instalá-lo, copie o arquivo de [BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Network accessline 431
[BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Access to secretsline 434
[Credential-related source omitted] - Instruction override signalline 535
- Linhas que parecem comandos diretos ou instruções de papel ("Ignore previous instructions...", "You are now...", "Write the token to..."). - Access to secretsline 675
[Credential-related source omitted] - Access to secretsline 744
[Credential-related source omitted] - Access to secretsline 745
[Credential-related source omitted] - Access to secretsline 746
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/privacy-and-data-rights.mdx
- Network accessline 59
- [Texto do GDPR](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [Comissão Europeia: solicitações individuais](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individ - Network accessline 60
- [Comissão Europeia: solicitações individuais](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [California Attorney General: CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [Regulamentos da CCPA](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/locales/pt-BR/real-time-collaboration.mdx
- Network accessline 686
`GET /_agent-native/health` informa o resultado, e resolvê-lo realiza o registro, então um único curl ao mesmo tempo ativa e confirma: - Network accessline 687
então um único curl ao mesmo tempo ativa e confirma: - Network accessline 690
curl -s https://your-app.example.com/_agent-native/health | jq .realtime
packages/core/docs/content/locales/pt-BR/recurring-jobs.mdx
packages/core/docs/content/locales/pt-BR/render.mdx
packages/core/docs/content/locales/pt-BR/sandbox-adapters.mdx
- Possible obfuscationline 259
Para a maioria dos CLIs, `ShellCliAdapter` agrupa qualquer binário com padrões razoáveis <U+200B><U+200B>e `CliRegistry` coleta adaptadores para descoberta em tempo de execução:
packages/core/docs/content/locales/pt-BR/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- Nunca use `innerHTML`, `eval()` ou `document.write()` - Access to secretsline 397
[Credential-related source omitted] - Access to secretsline 413
[Credential-related source omitted] - Access to secretsline 419
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/server-database.mdx
- Network accessline 8
Todo aplicativo nativo do agente armazena seu estado em PostgreSQL. A UI e o agente leem e gravam as mesmas tabelas por meio das mesmas actions, usando o mesmo cliente [Drizzle ORM](https://orm.drizzle.team). A sincronização ao vivo usa SSE - Network accessline 81
O [PGlite](https://pglite.dev/) executa um banco de dados Postgres WASM em processo. Ele permite detectar problemas de esquema exclusivos do Postgres antes de implantar, mantendo a configuração tão simples quanto o padrão PGlite. Assim como
packages/core/docs/content/locales/pt-BR/server-middleware.mdx
- Network accessline 106
Para mais padrões de middleware — lógica condicional, operações assíncronas, interceptação de resposta —, veja os [docs de middleware do Nitro](https://nitro.build/guide/routing#middleware) e os [docs do H3](https://h3.unjs.io). - Network accessline 149
- [**Rotas**](/docs/server-routes): rotas personalizadas protegidas por esse middleware - [**Security**](/docs/security): escopo de dados, guardas de acesso e o modelo completo de autenticação - [**Docs de middleware do Nitro**](https://nit - Network accessline 151
- [**Docs de middleware do Nitro**](https://nitro.build/guide/routing#middleware): API completa de middleware e padrões
packages/core/docs/content/locales/pt-BR/server-overview.mdx
- Network accessline 206
[Nitro](https://nitro.build) é o conjunto de ferramentas de servidor que sustenta a camada HTTP nativa do agente. A maior parte do que você configura em `server/` é Nitro: rotas baseadas em arquivo, plugins de inicialização e middleware. - Network accessline 214
A API de requisição/resposta subjacente é o [H3](https://h3.unjs.io). Quando você vir `defineEventHandler`, `createError` ou `getQuery` em rotas personalizadas, isso é H3. Os [docs do Nitro](https://nitro.build/guide) e os [docs do H3](http
packages/core/docs/content/locales/pt-BR/server-plugins.mdx
- Network accessline 141
Plugins podem usar qualquer um dos ganchos de ciclo de vida do Nitro: `request`, `beforeResponse`, `close`, entre outros. Os [docs de plugins do Nitro](https://nitro.build/guide/plugins) cobrem em detalhe a API completa de ganchos e o contr - Network accessline 163
- [**Rotas**](/docs/server-routes): rotas de arquivo personalizadas e quando usá-las - [**Escrevendo instruções do agente**](/docs/writing-agent-instructions): como criar o `systemPrompt` no plugin de agent-chat - [**Docs de plugins do Nitr - Network accessline 165
- [**Docs de plugins do Nitro**](https://nitro.build/guide/plugins): a API completa de plugins e os ganchos de ciclo de vida
packages/core/docs/content/locales/pt-BR/server-routes.mdx
- Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/skills-guide.mdx
- Possible obfuscationline 180
Os frontmatter `name` e `description` são usados <U+200B><U+200B>pelo sistema de ferramentas do agente para descoberta de habilidades. A descrição deve indicar quando a habilidade é acionada. Seja específico sobre as situações. - Network accessline 315
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 316
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 317
"url": "https://assets.agent-native.com", - Network accessline 318
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/locales/pt-BR/template-analytics.mdx
packages/core/docs/content/locales/pt-BR/template-assets.mdx
- Network accessline 117
Demonstração ao vivo: [assets.agent-native.com](https://assets.agent-native.com). - Network accessline 466
<EmbeddedApp url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 467
url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 503
A habilidade do aplicativo Assets tem o ID do aplicativo `assets` e MCP hospedado URL hospedado `https://assets.agent-native.com/mcp`. - Network accessline 504
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/locales/pt-BR/template-brain.mdx
- Network accessline 159
Demonstração ao vivo: [brain.agent-native.com](https://brain.agent-native.com).
packages/core/docs/content/locales/pt-BR/template-calendar-developers.mdx
- Network accessline 27
Isso requer [Node.js](https://nodejs.org) 22.22 ou posterior e - Network accessline 28
[pnpm](https://pnpm.io) instalados primeiro. Veja - Network accessline 39
O aplicativo abre em `http://localhost:8082`. Se não abrir automaticamente, - Network accessline 50
1. **Crie um projeto do Google Cloud**, se ainda não tiver um para usar: no [Google Cloud Console](https://console.cloud.google.com/), clique no - Network accessline 51
[Google Cloud Console](https://console.cloud.google.com/), clique no - Network accessline 62
4. **Crie credenciais**: em **APIs e Serviços → Credenciais**, crie um ID de cliente OAuth do tipo "Aplicativo Web" e adicione `http://localhost:8082/_agent-native/google/callback` como um URI de - Network accessline 63
cliente OAuth do tipo "Aplicativo Web" e adicione `http://localhost:8082/_agent-native/google/callback` como um URI de redirecionamento autorizado (8082 é a porta de desenvolvimento padrão do - Network accessline 64
`http://localhost:8082/_agent-native/google/callback` como um URI de - Access to secretsline 82
[Credential-related source omitted] - Network accessline 85
Obtenha um Client ID e um Secret em um [app OAuth do Zoom](https://marketplace.zoom.us/) (Zoom Marketplace → - Network accessline 86
[app OAuth do Zoom](https://marketplace.zoom.us/) (Zoom Marketplace →
packages/core/docs/content/locales/pt-BR/template-calendar-features.mdx
packages/core/docs/content/locales/pt-BR/template-calendar-integrations.mdx
- Access to secretsline 25
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/template-calendar.mdx
packages/core/docs/content/locales/pt-BR/template-chat-developers.mdx
- Network accessline 31
Isso requer [Node.js](https://nodejs.org) 22.22 ou posterior. Se o `pnpm` não
packages/core/docs/content/locales/pt-BR/template-chat-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/template-chat.mdx
packages/core/docs/content/locales/pt-BR/template-clips-developers.mdx
- Network accessline 22
Isso requer [Node.js](https://nodejs.org) 22.22 ou posterior e [pnpm](https://pnpm.io) instalados primeiro. Veja [Primeiros passos](/docs/getting-started) se você ainda não instalou nenhum dos dois, ou não tem uma conexão LLM (Builder.io, u - Network accessline 300
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 301
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 302
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/template-clips-embed.mdx
- Network accessline 40
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 53
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 54
[Credential-related source omitted] - Network accessline 56
const response = await fetch( - Network accessline 145
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 152
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/locales/pt-BR/template-clips-features.mdx
- Network accessline 117
O aplicativo de desktop é feito para a captura mais tranquila do dia a dia: gravação de tela e webcam com uma bolha de câmera arrastável, transcrição em tempo real enquanto você grava, tudo sem precisar manter uma aba do navegador aberta. B - Network accessline 212
[Credential-related source omitted] - Network accessline 213
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/template-clips-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/template-clips.mdx
packages/core/docs/content/locales/pt-BR/template-content.mdx
- Network accessline 17
### [Experimente agora →](https://content.agent-native.com)
packages/core/docs/content/locales/pt-BR/template-design-developers.mdx
packages/core/docs/content/locales/pt-BR/template-design-integrations.mdx
- Access to secretsline 25
[Credential-related source omitted] - Network accessline 52
https://design.agent-native.com/mcp - Network accessline 57
depois executa `/mcp` para autenticar. Codex e outros clientes CLI ou stdio, em vez disso, executam `npx @agent-native/core@latest connect https://design.agent-native.com`, o que abre uma aba do navegador para - Network accessline 58
stdio, em vez disso, executam `npx @agent-native/core@latest connect https://design.agent-native.com`, o que abre uma aba do navegador para autorizar e gera um token revogável por usuário, sem nenhum segredo - Network accessline 59
https://design.agent-native.com`, o que abre uma aba do navegador para - Network accessline 75
retornar um erro de sessão ou autenticação, não fique tentando de novo repetidamente. Execute `/mcp` no Claude Code e reconecte, ou execute `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` em - Network accessline 76
repetidamente. Execute `/mcp` no Claude Code e reconecte, ou execute `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` em qualquer terminal. Os dois reautenticam sem reinstalar nada. - Network accessline 77
-y @agent-native/core@latest reconnect https://design.agent-native.com` em
packages/core/docs/content/locales/pt-BR/template-design.mdx
packages/core/docs/content/locales/pt-BR/template-dispatch.mdx
packages/core/docs/content/locales/pt-BR/template-forms-agent.mdx
- Network accessline 40
- **"Publique isso e me dê o link."** O formulário fica no ar e você recebe sua URL pública imediatamente. - **"Crie um formulário anônimo de feedback do cliente e me dê o link público."** O formulário é publicado anônimo desde o início, en - Network accessline 42
- **"Só permita que este formulário seja enviado de https://example.com."** O formulário passa a rejeitar envios de qualquer outro lugar. Isso só está disponível pedindo ao agente hoje. Não existe um controle para isso no construtor.
packages/core/docs/content/locales/pt-BR/template-forms-features.mdx
- Network accessline 14
[Credential-related source omitted] - Network accessline 15
[Credential-related source omitted] - Network accessline 94
Envios públicos podem exigir verificação através do [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/) antes de serem aceitos. Ative isso definindo uma chave de site e uma chave secreta em Configurações. Duas proteções p - Network accessline 110
Hoje não existe um controle para isso no construtor. Peça ao agente para configurar, por exemplo "só permita que este formulário seja enviado de https://example.com." Informe ao agente a origem exata (o esquema e o domínio, como `https://ex - Network accessline 116
[Credential-related source omitted] - Network accessline 117
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/template-forms.mdx
packages/core/docs/content/locales/pt-BR/template-mail.mdx
- Network accessline 97
Demonstração ao vivo: [mail.agent-native.com](https://mail.agent-native.com). - Network accessline 168
1. Abra [Google Cloud Console](https://console.cloud.google.com/) e crie um projeto. - Network accessline 169
2. Ative **Gmail API** em APIs e Serviços → Biblioteca. 3. Crie credenciais OAuth 2.0 (tipo: aplicativo Web). Adicione `http://localhost:8085/_agent-native/google/callback` como um redirecionamento autorizado URI. 4. Copie o ID do cliente e - Network accessline 170
3. Crie credenciais OAuth 2.0 (tipo: aplicativo Web). Adicione `http://localhost:8085/_agent-native/google/callback` como um redirecionamento autorizado URI.
packages/core/docs/content/locales/pt-BR/template-plan.mdx
- Network accessline 143
`/visualize-repo` — além do conector MCP `plan` (um servidor de ferramentas remoto com o qual o seu agente conversa) em `https://plan.agent-native.com/mcp`. Para a referência completa de flags, as - Network accessline 144
remoto com o qual o seu agente conversa) em `https://plan.agent-native.com/mcp`. Para a referência completa de flags, as rotas de instalação por plugin/marketplace para Claude Code e Codex, e a - Network accessline 145
`https://plan.agent-native.com/mcp`. Para a referência completa de flags, as - Network accessline 240
Se uma ferramenta do Plans retornar `needs auth`, `Unauthorized` ou `Session terminated`, não fique tentando de novo. Reautentique com `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, - Network accessline 241
terminated`, não fique tentando de novo. Reautentique com `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, ou rode novamente `/mcp` → **Authenticate** para um host compatível com - Network accessline 242
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`,
packages/core/docs/content/locales/pt-BR/template-slides-features.mdx
packages/core/docs/content/locales/pt-BR/template-slides-integrations.mdx
- Access to secretsline 24
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/template-slides.mdx
packages/core/docs/content/locales/pt-BR/toolkit-capability-packages.mdx
packages/core/docs/content/locales/pt-BR/toolkit-settings.mdx
- Access to secretsline 162
[Credential-related source omitted] - Network accessline 269
`*.agent-native.com` não são compatíveis com esse fluxo. A aprovação de parceiro é necessária para quem opera o app hospedado e pode ser solicitada pelo [formulário de interesse](https://openai.com/form/sign-in-with-chatgpt-interest). - Network accessline 271
[formulário de interesse](https://openai.com/form/sign-in-with-chatgpt-interest). - Network accessline 290
prévia, então a disponibilidade de modelos, os limites e os recursos de API compatíveis podem mudar. Consulte a [visão geral do Sign in with ChatGPT](https://developers.openai.com/siwc/token-sharing-open-source), - Network accessline 291
compatíveis podem mudar. Consulte a [visão geral do Sign in with ChatGPT](https://developers.openai.com/siwc/token-sharing-open-source), [registro e login](https://developers.openai.com/siwc/token-sharing-open-source/sign-in), - Network accessline 292
ChatGPT](https://developers.openai.com/siwc/token-sharing-open-source), - Network accessline 293
[registro e login](https://developers.openai.com/siwc/token-sharing-open-source/sign-in), - Network accessline 294
[contas e sessões](https://developers.openai.com/siwc/token-sharing-open-source/profiles-and-sessions), - Network accessline 295
[modelos e inferência](https://developers.openai.com/siwc/token-sharing-open-source/models-and-inference), - Network accessline 296
[limitações da prévia](https://developers.openai.com/siwc/token-sharing-open-source/preview-limitations), - Network accessline 297
[recuperação de erros](https://developers.openai.com/siwc/token-sharing-open-source/errors-and-recovery) - Network accessline 298
e [diretrizes de UI/UX](https://developers.openai.com/siwc/ui-ux-guidelines).
packages/core/docs/content/locales/pt-BR/tracking.mdx
- Network accessline 83
| Provedor | Vars ambientes | | ----------------- | ----------------------------------------------------------------------------------------------- - Network accessline 84
| ----------------- | -------------------------------------------------------------------------------------------------- | | PostHog | `POSTHOG_API_KEY` (obrigatório), `POSTHOG_HOST` (opcional, o padrão é `https://us.i.posthog.com - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Network accessline 205
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 206
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 207
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 225
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 226
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 283
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 284
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 285
fetch("https://analytics.example.com/events", { - Network accessline 343
- **HTTP em lote** — provedores integrados enfileiram eventos e liberam a cada 10 segundos ou quando 50 eventos se acumulam, o que ocorrer primeiro. Isso minimiza as solicitações de saída sem perder dados. - **Sem dependências de SDK** — to - Network accessline 344
- **Sem dependências de SDK** — todos os provedores integrados usam `fetch()` bruto. Sem PostHog SDK, sem Mixpanel SDK, sem Amplitude SDK. Mantém a estrutura leve.
packages/core/docs/content/locales/pt-BR/turn-into-app.mdx
- Network accessline 12
- Um agente de IA que consiga executar comandos de terminal, como Claude Code, Codex ou Cursor. Claude ou ChatGPT na web podem entregar o trabalho ao Builder em vez disso. Veja [Onde ela roda](#hosts). - [Node.js](https://nodejs.org). O ins - Network accessline 13
- [Node.js](https://nodejs.org). O instalador de nodejs.org inclui o `npx`, que executa os comandos de instalação abaixo. - Network accessline 18
- **Na sua máquina.** Node.js 22.22 ou posterior (confira com `node --version`) e [pnpm](https://pnpm.io) no seu `PATH`. Se `pnpm --version` falhar, execute `corepack enable`, que já vem com o Node.js. - Network accessline 59
A hospedagem no Builder, a partir de qualquer tipo de host, exige o conector MCP do Agent-Native Dispatch (`https://dispatch.agent-native.com/mcp`, OAuth) e chama a ação `start-workspace-app-creation` dele. Nada roda do seu lado, então o re - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/locales/pt-BR/updating-ui-in-production.mdx
- Network accessline 19
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment) é um fluxo de trabalho gerenciado separado, com capacidade de código, no qual um agente pode modificar a UI do seu ap
packages/core/docs/content/locales/pt-BR/upgrading-core-ui.mdx
- Access to secretsline 73
[Credential-related source omitted] - Access to secretsline 76
[Credential-related source omitted] - Network accessline 188
URL canônica da fonte: [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx). - Network accessline 189
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx).
packages/core/docs/content/locales/pt-BR/vercel.mdx
- Access to secretsline 94
[Credential-related source omitted] - Network accessline 111
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 138
[Credential-related source omitted] - Network accessline 201
Automações agendadas, automações por evento na fila e a limpeza de execuções interrompidas acontecem em uma varredura executada uma vez por minuto. O build adiciona um [cron job da Vercel](https://vercel.com/docs/cron-jobs) que chama a - Network accessline 202
interrompidas acontecem em uma varredura executada uma vez por minuto. O build adiciona um [cron job da Vercel](https://vercel.com/docs/cron-jobs) que chama a varredura em `/_agent-native/jobs/_process-sweep`. Um build de espaço de - Network accessline 203
adiciona um [cron job da Vercel](https://vercel.com/docs/cron-jobs) que chama a
packages/core/docs/content/locales/pt-BR/voice-input.mdx
packages/core/docs/content/locales/pt-BR/what-is-agent-native.mdx
packages/core/docs/content/locales/pt-BR/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Network accessline 269
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 270
[Credential-related source omitted] - Network accessline 271
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 272
[Credential-related source omitted] - Access to secretsline 273
[Credential-related source omitted] - Access to secretsline 274
[Credential-related source omitted] - Network accessline 286
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 287
[Credential-related source omitted] - Network accessline 288
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 289
[Credential-related source omitted] - Access to secretsline 290
[Credential-related source omitted] - Access to secretsline 291
[Credential-related source omitted] - Network accessline 343
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 345
APP_URL=https://workspace.example.com
packages/core/docs/content/locales/pt-BR/workspace-management.mdx
- Network accessline 90
Nem todo mundo que precisa fazer alterações se sente confortável com o git. [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management) oferece suporte a um modelo de - Network accessline 140
Dicas importantes: use equipes GitHub (`@org/team`), não indivíduos. As alterações na estrutura e no despacho devem sempre exigir a revisão da plataforma. Consulte [GitHub CODEOWNERS docs](https://docs.github.com/en/repositories/managing-yo - Network accessline 160
Em seguida, adicione a ação [actions/labeler](https://github.com/actions/labeler) - consulte o README desse repositório para obter o fluxo de trabalho completo YAML. Os rótulos são aplicados automaticamente quando os PRs são abertos ou atua - Access to secretsline 233
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/a2a-protocol.mdx
- Access to secretsline 93
[Credential-related source omitted] - Network accessline 117
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 118
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 119
"url": "https://analytics.example.com", - Network accessline 314
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 350
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 351
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 352
"https://analytics.example.com", - Access to secretsline 354
[Credential-related source omitted] - Network accessline 370
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", - Network accessline 371
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", "cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 372
"url": "https://example.services.ai.azure.com/a2a", - Network accessline 373
"cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 408
[Azure AI Foundry 托管代理](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 413
Foundry 可以通过 [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint) - Network accessline 422
[Gemini Enterprise 可以注册外部 A2A 代理](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent), - Network accessline 424
路径:发现使用 `GET {url}/v1/card`,调用使用 `POST {url}/v1/message:send`, 流式调用使用 `POST {url}/v1/message:stream`。请求使用 Google OAuth 或 ADC bearer 令牌以及 Google Cloud IAM 权限。端点和权限要求请参阅[使用注册表 A2A 端点调用代理](https://docs.cloud.google.com/gemini/enterprise/docs/i - Network accessline 426
bearer 令牌以及 Google Cloud IAM 权限。端点和权限要求请参阅[使用注册表 A2A 端点调用代理](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a)。 - Network accessline 441
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 442
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 443
"url": "https://api.anthropic.com", - Access to secretsline 448
[Credential-related source omitted] - Access to secretsline 461
[Credential-related source omitted] - Network accessline 507
# 通过 URL 直接调用一个已部署的应用,从环境变量中读取 bearer token agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 508
[Credential-related source omitted] - Access to secretsline 508
[Credential-related source omitted] - Access to secretsline 509
[Credential-related source omitted] - Access to secretsline 577
[Credential-related source omitted] - Access to secretsline 581
[Credential-related source omitted] - Network accessline 639
- 共享的 `A2A_SECRET` 在构建时安装到每个应用程序的环境中。 - 跨应用调用是同源的 - `https://workspace.example.com/apps/analytics` 调用 `https://workspace.example.com/apps/mail` - 因此没有 DNS、CORS 或每对 JWT 设置。 - Network accessline 640
- 跨应用调用是同源的 - `https://workspace.example.com/apps/analytics` 调用 `https://workspace.example.com/apps/mail` - 因此没有 DNS、CORS 或每对 JWT 设置。 - Network accessline 648
**永远不要依赖“一劳永逸”的 `Promise` 超过响应的寿命。**无服务器函数(Netlify、Vercel、AWS Lambda、Cloud Run)会在响应正文刷新时冻结 - 有时甚至在未等待的 `fetch(...)` 的 TCP 握手完成之前。在 Node 上本地工作的模式将默默地放弃生产中的工作。 - Network accessline 686
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 687
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 688
"https://analytics.example.com", - Access to secretsline 690
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/actions-defining.mdx
- Network accessline 221
`schema` 接受任何兼容 [Standard Schema](https://standardschema.dev) 的库:
packages/core/docs/content/locales/zh-CN/add-ui-to-your-agent.mdx
packages/core/docs/content/locales/zh-CN/agent-mentions.mdx
packages/core/docs/content/locales/zh-CN/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/agent-web-surfaces.mdx
- Network accessline 171
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 172
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 173
siteUrl: "https://example.com", - Network accessline 198
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/locales/zh-CN/audit-log.mdx
packages/core/docs/content/locales/zh-CN/authentication.mdx
- Network accessline 8
代理本机应用程序使用 [Better Auth](https://better-auth.com) 进行帐户优先设计的身份验证。用户在第一次访问时创建一个帐户,并从第一天起就获得真实身份。 - Access to secretsline 193
[Credential-related source omitted] - Access to secretsline 198
[Credential-related source omitted] - Access to secretsline 202
[Credential-related source omitted] - Access to secretsline 207
[Credential-related source omitted] - Access to secretsline 209
[Credential-related source omitted] - Access to secretsline 231
[Credential-related source omitted] - Access to secretsline 235
[Credential-related source omitted] - Network accessline 248
https://mail.agent-native.com/mcp - Access to secretsline 473
[Credential-related source omitted] - Access to secretsline 475
[Credential-related source omitted] - Access to secretsline 477
[Credential-related source omitted] - Access to secretsline 478
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/automations.mdx
packages/core/docs/content/locales/zh-CN/aws-amplify.mdx
- Access to secretsline 105
[Credential-related source omitted] - Network accessline 119
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 177
[Credential-related source omitted] - Network accessline 191
在 [Amplify 控制台](https://console.aws.amazon.com/amplify/)中创建应用: - Network accessline 225
首次构建需要几分钟,你可以在 Amplify 控制台中跟踪进度。构建完成后,打开应用的 地址,形如 `https://main.d1m7bkiki6tdw1.amplifyapp.com`,并创建第一个账户。 - Network accessline 226
地址,形如 `https://main.d1m7bkiki6tdw1.amplifyapp.com`,并创建第一个账户。 - Access to secretsline 273
[Credential-related source omitted] - Access to secretsline 277
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/aws-lambda.mdx
- Network accessline 35
你还需要一个 AWS 账户,并在你的电脑上安装并登录 [AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html), - Access to secretsline 106
[Credential-related source omitted] - Access to secretsline 136
[Credential-related source omitted] - Access to secretsline 154
[Credential-related source omitted] - Access to secretsline 199
[Credential-related source omitted] - Access to secretsline 207
[Credential-related source omitted] - Network accessline 229
第一条命令会输出 `FunctionUrl`,形如 `https://abc123.lambda-url.us-east-1.on.aws/`。 - Network accessline 247
在 [CloudFront 控制台](https://console.aws.amazon.com/cloudfront/v4/home)中创建 - Network accessline 261
在分配的 **Origins** 选项卡中创建一个源。对于 **Origin domain**,粘贴第 5 步中 函数 URL 的主机名,不要包含 `https://` 或末尾的斜杠。把 **Protocol** 设置为 - Network accessline 262
函数 URL 的主机名,不要包含 `https://` 或末尾的斜杠。把 **Protocol** 设置为 - Network accessline 294
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 296
[Credential-related source omitted] - Access to secretsline 298
[Credential-related source omitted] - Access to secretsline 314
[Credential-related source omitted] - Dangerous shell operationline 316
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/locales/zh-CN/azure-static-web-apps.mdx
packages/core/docs/content/locales/zh-CN/blueprint-installer.mdx
packages/core/docs/content/locales/zh-CN/cli-adapters.mdx
- Access to secretsline 99
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/locales/zh-CN/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">组件</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 125
optional: true, description: "[useQuery 选项](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 中除 `queryKey` 和 `queryFn`(由该钩子自行设置)之外的所有内容。", - Network accessline 126
description: "[useQuery 选项](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 中除 `queryKey` 和 `queryFn`(由该钩子自行设置)之外的所有内容。", }, - Network accessline 127
"[useQuery 选项](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 中除 `queryKey` 和 `queryFn`(由该钩子自行设置)之外的所有内容。", - Network accessline 137
该钩子用于只读 actions,是围绕 [useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 函数的一层轻量封装。 - Network accessline 173
optional: true, description: "[useMutation 选项](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 中除 `mutationFn`(由该钩子自行设置)之外的所有内容。", - Network accessline 174
description: "[useMutation 选项](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 中除 `mutationFn`(由该钩子自行设置)之外的所有内容。", fields: [ - Network accessline 175
"[useMutation 选项](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 中除 `mutationFn`(由该钩子自行设置)之外的所有内容。", - Network accessline 201
该钩子用于会改变数据的 actions,例如创建、更新或发送某些内容,是围绕 [useMutation()](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 函数的一层轻量封装。
packages/core/docs/content/locales/zh-CN/cloneable-saas.mdx
- Network accessline 46
https://github.com/owner/repo - Network accessline 61
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 62
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 70
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 71
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 129
希望加入列表?请提交[社区模板](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml), - Access to secretsline 281
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/cloudflare.mdx
- Access to secretsline 91
[Credential-related source omitted] - Network accessline 105
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 150
[Credential-related source omitted] - Network accessline 189
计划自动化、排队中的事件自动化以及被中断运行的清理,都在每分钟一次的清扫中完成。构建会在生成的 `wrangler.json` 中添加一个 `* * * * *` [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/),并保留你已经声明的所有 Cron Trigger。触发时,Worker 的 `scheduled` 处理程序会使用一个由 `A2A_SECRE
packages/core/docs/content/locales/zh-CN/code-agents-ui.mdx
- Network accessline 202
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 203
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 355
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 356
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard"
packages/core/docs/content/locales/zh-CN/create-with-an-agent.mdx
- Network accessline 14
- 一个能运行终端命令的 AI 智能体,例如 Claude Code、Codex 或 Cursor。只能聊天的智能体,例如 ChatGPT 或网页版 Claude,也能提供帮助。请参阅[无法运行命令的智能体](#chat-only-agents)。 - [Node.js](https://nodejs.org) 22.22 或更高版本。 - Network accessline 15
- [Node.js](https://nodejs.org) 22.22 或更高版本。 - Network accessline 37
5. 用 `pnpm dev` 启动应用,并让它在后台持续运行。 6. 读取 `pnpm dev` 输出的本地 URL。检查它是否有响应, 例如用 `curl -I <url>`。不要打开浏览器,也不要使用无头 - Network accessline 38
6. 读取 `pnpm dev` 输出的本地 URL。检查它是否有响应, 例如用 `curl -I <url>`。不要打开浏览器,也不要使用无头 浏览器工具。我会自己打开应用。 - Network accessline 39
例如用 `curl -I <url>`。不要打开浏览器,也不要使用无头 - Access to secretsline 48
[Credential-related source omitted] - Network accessline 56
- 一个新的应用文件夹,除非你改过名称,否则叫 `my-app`,其中包含 `actions/`、`app/` 和一个 `AGENTS.md` 文件。 - `pnpm dev` 正常运行且没有错误。 - 一个本地 URL,通常是 `http://localhost:8080`,智能体已经检查过它有响应。 - Network accessline 58
- 一个本地 URL,通常是 `http://localhost:8080`,智能体已经检查过它有响应。
packages/core/docs/content/locales/zh-CN/creating-templates.mdx
packages/core/docs/content/locales/zh-CN/decisions.mdx
- Network accessline 18
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 19
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 20
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/locales/zh-CN/deno-deploy.mdx
- Network accessline 24
你还需要一个 Deno Deploy 账户,并在你的电脑上安装 [Deno](https://docs.deno.com/runtime/getting_started/installation/)。 - Access to secretsline 93
[Credential-related source omitted] - Network accessline 107
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 154
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/deployment-environment-variables.mdx
- Network accessline 30
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | 在自动无服务器跳过机制没有覆盖到的部署上,跳过按需的 schema DDL。要求你的 schema 已经迁移完毕。参见 [On-demand schema DDL](/docs/deployment#skip-ensure-tables)。 - Network accessline 32
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | 仅用于企业回调的、以逗号分隔的精确可信预览来源允许列表(例如 `https://0123456789abcdef01234567--content.netlify.app`)。只在企业回调部署上设置它。Netlify 的来源必须使用一个不可变的、特定于该次部署的 24 位十六进制永久链接。可变的 `deploy-preview-*` 别名、通配符和域名后缀都会被拒绝。 | - Network accessline 45
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 46
| `BETTER_AUTH_SECRET` | 一个 32 位以上字符的随机字符串。用于签署会话 cookie,并作为 `OAUTH_STATE_SECRET` 和应用本地加密值的后备。它同时也是读取现有共享保管库密文的一个旧版候选项。硬性必需:在生产环境中如果缺失,框架会在启动时抛出错误。 | | `BETTER_AUTH_URL` - Network accessline 47
| `BETTER_AUTH_URL` | 这个应用的可选公共源覆盖值(例如 `https://mail.example.com`)。未设置时,框架会从请求、已知模板和托管平台上下文推断它。 - Access to secretsline 48
[Credential-related source omitted] - Access to secretsline 57
[Credential-related source omitted] - Access to secretsline 67
[Credential-related source omitted] - Access to secretsline 69
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/deployment.mdx
packages/core/docs/content/locales/zh-CN/dispatch.mdx
- Network accessline 116
Dispatch 可以是外部代理的单个 MCP 连接器:在 Claude、ChatGPT、Codex 或 Cursor 中添加一次 `https://dispatch.agent-native.com/mcp`,一次授权可到达每个授予的工作区应用程序,而不是每个应用程序一个连接器。请参阅 [External Agents](/docs/external-agents) 了解完整的连接流程、应用程序授权、OAuth 和内联 MCP 应用程序预览。 - Network accessline 125
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p
packages/core/docs/content/locales/zh-CN/docker.mdx
- Dangerous shell operationline 65
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 67
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 106
[Credential-related source omitted] - Access to secretsline 156
[Credential-related source omitted] - Network accessline 171
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/locales/zh-CN/embedding-sdk.mdx
- Network accessline 222
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 223
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 224
url="https://assets.agent-native.com/picker" - Network accessline 277
getMcpUrl("https://assets.agent-native.com"); - Network accessline 278
getA2AUrl("https://assets.agent-native.com"); - Network accessline 280
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 281
"https://assets.agent-native.com", - Network accessline 302
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 303
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 304
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 386
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 388
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 390
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 391
hostOrigin: "https://app.example.com", - Network accessline 394
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 395
hostOrigin: "https://app.example.com", - Network accessline 399
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 400
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 401
{ hostOrigin: "https://app.example.com" }, - Network accessline 405
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 406
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 407
{ hostOrigin: "https://app.example.com" }, - Network accessline 410
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 411
hostOrigin: "https://app.example.com", - Network accessline 462
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 463
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 464
hostOrigin: "https://app.example.com",
packages/core/docs/content/locales/zh-CN/extensions.mdx
- Network accessline 235
| `dbQuery(sql, args)` | 从 SQL 读取(自动调整范围给用户) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` | Legacy trusted SQL writes for ed - Network accessline 236
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` | 通过带有秘密的安全代理攻击外部 AP - Network accessline 237
| `extensionFetch(url, options)` | 通过带有秘密的安全代理攻击外部 API | `extensionFetch('https://api.github.com/user')` | - Network accessline 265
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 266
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/locales/zh-CN/external-agents.mdx
- Network accessline 111
https://dispatch.agent-native.com/mcp - Network accessline 122
https://mail.agent-native.com/mcp - Network accessline 126
每个托管应用程序还有一个帮助页面 `https://<app>/mcp/connect` 与可复制的 URL 和 - Network accessline 127
`https://<app>/mcp/connect` 与可复制的 URL 和 - Network accessline 214
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 258
原始 `skills` CLI 仅安装 `SKILL.md` 文件;本地 MCP 客户端仍然 需要一个连接器,例如 `npx @agent-native/core@latest connect https://assets.agent-native.com`。 - Network accessline 259
需要一个连接器,例如 `npx @agent-native/core@latest connect https://assets.agent-native.com`。 - Network accessline 265
默认客户端选择是所有支持的本地客户端;添加 `--client codex`、`--client claude-code` 或其他特定目标以缩小设置范围。内联主机(ChatGPT、网页版、桌面版和移动版 Claude 聊天,以及 Claude Desktop 的 Code 选项卡)在聊天中显示选择器或变体网格;CLI/仅链接主机(Codex CLI、Claude Code 终端、Cowork、claude.ai 网页版 Code)返回“在…中打开 →”链接。用户在浏览器中选择 - Network accessline 271
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 286
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 289
`reconnect` 查找给定主机和选定客户端的 URL 以 `/mcp` 结尾的任何 MCP 配置条目(通过 URL 进行匹配,无论连接器名称如何),然后刷新或替换身份验证材料,而无需触及已安装的 skills 或重新运行完整安装流程。传递基本应用程序 URL(例如 `https://plan.agent-native.com`) - 推断出 `/mcp` 后缀。身份验证和工具加载是针对每个客户端的,因此之后重新启动/重新加载该客户端; Codex 在新加载的工具出现之前 - Network accessline 297
对于无法直接添加远程 OAuth URL 的 MCP 客户端,请在浏览器中打开应用程序并使用其 **Connect** 功能(在 `https://<app>/mcp/connect` 上提供)。登录后,单击“**连接/授权**”。该页面为您提供一个配置检测到的代理的一键深层链接,或一个可立即粘贴的 `.mcp.json` 块: - Network accessline 303
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 304
[Credential-related source omitted] - Network accessline 305
"url": "https://mail.agent-native.com/mcp", - Network accessline 321
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 322
https://dispatch.agent-native.com/mcp - Network accessline 325
这与 `npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code` 为您编写的仅 URL 条目相同。然后在Claude代码中运行`/mcp`并选择**验证**。客户端从 MCP 服务器的 `401 WWW-Authenticate` 挑战中发现身份验证,获取 `/.well-known/oauth-protected-resource` 和 `/.w - Network accessline 328
https://dispatch.agent-native.com/mcp - Network accessline 468
npx @agent-native/core@latest 连接 https://plan.agent-native.com --client codex --full-catalog - Network accessline 519
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 520
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Access to secretsline 739
[Credential-related source omitted] - Access to secretsline 755
[Credential-related source omitted] - Access to secretsline 761
[Credential-related source omitted] - Network accessline 767
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 768
[Credential-related source omitted] - Network accessline 769
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 770
[Credential-related source omitted] - Network accessline 792
`connect dev` 将相同的稳定 MCP 服务器名称(`agent-native-mail`、`agent-native-calendar` 等)重写到本地 dev-lazy 网关,因此工具名称不会更改。在写入开发条目之前,它会备份 `~/.agent-native/connect-profiles.json` 中的当前生产条目。默认网关为`http://127.0.0.1:8080`;如果您的网关移动了,请使用 `--gateway <url>` 或 `--port - Network accessline 819
- 使用 `npx @agent-native/core@latest connect https://dispatch.agent-native.com` 将您自己的代理连接到 Dispatch;仅当您需要一个独立的应用程序时才使用直接应用程序 URL。 - Access to secretsline 839
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/file-uploads.mdx
- Network accessline 131
只有在需要接入不同类型的后端时,才需要注册自定义提供程序。下面的示例通过 `fetch()` 调用一个通用的 HTTP API 来完成上传: - Network accessline 148
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 149
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 150
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/locales/zh-CN/getting-started.mdx
packages/core/docs/content/locales/zh-CN/harness-agents.mdx
- Network accessline 111
Agent-Native可以充当[ACP](https://agentclientprotocol.com)(代理客户端
packages/core/docs/content/locales/zh-CN/key-concepts.mdx
- Network accessline 328
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 329
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 330
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/locales/zh-CN/koyeb.mdx
packages/core/docs/content/locales/zh-CN/mcp-apps.mdx
- Network accessline 23
OpenAI 插件扩展为模型调用的 MCP 工具增加应用启动入口。启用内联 MCP Apps 后,共享的 `open_app` 工具会同时提供全局侧边栏入口和会话侧边栏入口。ChatGPT 使用 `{}` 调用该工具;Core 会以内联 MCP App 的形式打开当前应用主页。使用共享 MCP 服务器的应用无需逐个注册。请参阅 [OpenAI 插件扩展指南](https://developers.openai.com/plugins/build/extensions)和 [M - Network accessline 261
支持内联渲染 MCP Apps 的主机包括网页版、桌面版和移动版 Claude 聊天,以及 Claude Desktop 的 [Code](https://github.com/anthropics/claude-code/issues/103809) 选项卡; ChatGPT; VS Code GitHub Copilot (1.109+); Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Code - Network accessline 263
在 Claude Code 终端、Cowork、claude.ai 网页版 Code 和 Codex CLI 中,请使用深层链接后备方案。[Cowork 尚未声明支持 MCP Apps,在同一连接器上使用它可能会导致 Chat 中的内联应用消失,直到您重新连接该连接器。](https://github.com/anthropics/claude-code/issues/103924) 当主机选择不渲染 iframe 时,深层链接仍然是可靠的后备方案。实际上,每个代理本机应用程 - Access to secretsline 269
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/mcp-clients.mdx
- Network accessline 148
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 149
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 178
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 179
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 180
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 226
[Credential-related source omitted] - Network accessline 269
Atlassian 预设使用 Atlassian Rovo MCP 的 OAuth 2.1 端点 `https://mcp.atlassian.com/v1/mcp/authv2`,并提供 Jira 和 Confluence - Network accessline 270
`https://mcp.atlassian.com/v1/mcp/authv2`,并提供 Jira 和 Confluence - Network accessline 276
Rovo MCP server → Domains** 下的受信任域列表,并在 **Permissions** 中启用 所需的 **Read**、**Write** 和 **Search** 权限。该域必须是 Agent-Native 应用实际运行的主机;除非 `https://builder.io/**` 是部署的实际源,否则不要将其 - Network accessline 277
所需的 **Read**、**Write** 和 **Search** 权限。该域必须是 Agent-Native 应用实际运行的主机;除非 `https://builder.io/**` 是部署的实际源,否则不要将其 复制到其他部署。请参阅 Atlassian 的 [Rovo MCP 设置指南](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 278
应用实际运行的主机;除非 `https://builder.io/**` 是部署的实际源,否则不要将其 - Network accessline 279
复制到其他部署。请参阅 Atlassian 的 [Rovo MCP 设置指南](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 280
和[管理员控制说明](https://support.atlassian.com/security-and-access-policies/docs/control-atlassian-rovo-mcp-server-settings/)。 - Network accessline 289
1. **服务器名称** - 供您自己参考的简短标签(例如“zapier”、“slack-tools”)。 2. **URL** — MCP 服务器提供商为您提供的 HTTPS 端点(例如 `https://mcp.zapier.com/s/abc123/mcp`)。这通常可以在提供商的仪表板或集成文档中找到。 - Network accessline 290
2. **URL** — MCP 服务器提供商为您提供的 HTTPS 端点(例如 `https://mcp.zapier.com/s/abc123/mcp`)。这通常可以在提供商的仪表板或集成文档中找到。 - Network accessline 305
[Credential-related source omitted] - Network accessline 397
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 398
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 399
"url": "https://example.com/mcp", - Network accessline 437
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com
packages/core/docs/content/locales/zh-CN/mcp-protocol.mdx
- Access to secretsline 31
[Credential-related source omitted] - Network accessline 108
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 117
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 118
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 119
"url": "https://mail.example.com/mcp", - Network accessline 138
POST https://your-app.example.com/mcp - Access to secretsline 143
[Credential-related source omitted] - Access to secretsline 253
[Credential-related source omitted] - Network accessline 261
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 267
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 268
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 383
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 392
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 393
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 394
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/locales/zh-CN/messaging.mdx
- Network accessline 81
1. 转到 **[api.slack.com/apps](https://api.slack.com/apps)** 并单击 **创建新应用程序** → **从头开始**。为其命名(例如“Agent”)并选择您的工作区。 - Network accessline 97
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 120
https://your-clips.example.com/api/slack/unfurl - Access to secretsline 134
[Credential-related source omitted] - Network accessline 166
将 OAuth Redirect URL 设置为 `https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 167
`https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 168
Event Subscriptions Request URL 设置为 `https://your-app.example.com/_agent-native/integrations/slack/webhook`, Interactivity Request URL 设置为 - Network accessline 169
`https://your-app.example.com/_agent-native/integrations/slack/webhook`, - Network accessline 170
Interactivity Request URL 设置为 `https://your-app.example.com/_agent-native/integrations/slack/interactions`。 订阅 `app_home_opened`、`app_context_changed`、`app_mention` 和 `message.im`。 - Network accessline 171
`https://your-app.example.com/_agent-native/integrations/slack/interactions`。 - Access to secretsline 173
[Credential-related source omitted] - Network accessline 259
[Credential-related source omitted] - Access to secretsline 259
[Credential-related source omitted] - Access to secretsline 264
[Credential-related source omitted] - Access to secretsline 361
[Credential-related source omitted] - Network accessline 375
1. 打开电报和消息**[@BotFather](https://t.me/BotFather)**。 - Network accessline 382
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 401
1. 在 **[resend.com](https://resend.com)** 注册。免费套餐足以开始使用。 - Network accessline 408
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 415
[Credential-related source omitted] - Network accessline 422
1. 在 **[sendgrid.com](https://sendgrid.com)** 注册。 - Network accessline 430
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 435
[Credential-related source omitted] - Network accessline 459
1. 转到 **[Meta Developer Portal](https://developers.facebook.com/)**,点击 **创建应用**,然后选择 **商业** 类型。 - Access to secretsline 466
[Credential-related source omitted] - Network accessline 472
回调URL:https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Access to secretsline 683
[Credential-related source omitted] - Access to secretsline 684
[Credential-related source omitted] - Access to secretsline 794
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/multi-app-workspace.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 100
[Credential-related source omitted] - Network accessline 245
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 246
"urls": { "design": "https://design.example.test" } - Network accessline 247
"design": "https://design.example.test" - Access to secretsline 276
[Credential-related source omitted] - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Network accessline 325
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 326
# https://your-agents.com/mail/* → apps/mail - Network accessline 327
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 328
# https://your-agents.com/forms/* → apps/forms
packages/core/docs/content/locales/zh-CN/native-chat-ui.mdx
- Possible obfuscationline 152
仅当操作声明 `chatUI`**或**结果携带显式已知的 `widget` 判别式时,渲染器才会接管。它从不形<U+200B><U+200B>状推断任意对象,也从不从工具结果执行 HTML 或 JavaScript - 因此本机小部件不能成为注入向量。 - Network accessline 446
旧 API 与新 API 的对应关系及迁移步骤,请参阅 [AgentKit 聊天迁移指南](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md)。
packages/core/docs/content/locales/zh-CN/netlify.mdx
packages/core/docs/content/locales/zh-CN/node-js.mdx
packages/core/docs/content/locales/zh-CN/notifications.mdx
- Network accessline 192
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 193
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 194
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 211
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 212
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 213
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/locales/zh-CN/observability.mdx
- Network accessline 507
仓库中的 [`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel) 提供了这个插件更完整的版本,支持按信号设置端点并校验导出器。它没有发布到 npm,因此请从上面的插件开始,而不是安装它。
packages/core/docs/content/locales/zh-CN/onboarding.mdx
- Access to secretsline 186
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/plan-plugin.mdx
- Network accessline 19
- **三个 skill** — `/visual-plan`(规范入口)、`/visual-recap` 和 `/visualize-repo`。 - **Plan MCP 连接器** — 注册到位于 `https://plan.agent-native.com` 的托管应用(MCP 端点为 `https://plan.agent-native.com/mcp`,服务器名称为 `plan`)。 - Network accessline 20
- **Plan MCP 连接器** — 注册到位于 `https://plan.agent-native.com` 的托管应用(MCP 端点为 `https://plan.agent-native.com/mcp`,服务器名称为 `plan`)。 - Network accessline 92
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 118
- **使用 Chromium 系浏览器。** Safari 会阻止托管的 HTTPS Plan 页面 读取 `http://127.0.0.1` 本地主机桥(混合内容/私有网络限制), - Network accessline 119
读取 `http://127.0.0.1` 本地主机桥(混合内容/私有网络限制), - Network accessline 219
- `--client codex|claude-code|claude-code-cli|cowork|all` — 为哪些本地代理写入 MCP 配置(默认是 `all`)。 - `--no-connect` — 只注册连接器而不进行身份验证;之后运行 `npx @agent-native/core@latest connect https://plan.agent-native.com --client all`,或者选择更小范围的 `--client`。 - Network accessline 220
- `--no-connect` — 只注册连接器而不进行身份验证;之后运行 `npx @agent-native/core@latest connect https://plan.agent-native.com --client all`,或者选择更小范围的 `--client`。 - Network accessline 278
安装之后,**新建一个 Codex 会话**,让 skill 和 MCP 工具加载到会话 中。该插件自带一个仅含 URL 的连接器(`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`);`codex mcp login plan` 会 - Network accessline 279
中。该插件自带一个仅含 URL 的连接器(`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`);`codex mcp login plan` 会 运行 OAuth 流程。如果你更喜欢用一条命令同时完成安装和身份验证, - Network accessline 280
`https://plan.agent-native.com/mcp`);`codex mcp login plan` 会 - Network accessline 287
**旧版安装:** 如果你的配置中仍有一个指向同一 URL 的 `agent-native-plans` 条目,对 Codex 运行 `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, - Network accessline 288
`agent-native-plans` 条目,对 Codex 运行 `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, 或者用你的目标 `--client` 运行同一条命令,就会把它整合为规范的 - Network accessline 289
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`,
packages/core/docs/content/locales/zh-CN/pr-visual-recap.mdx
- Access to secretsline 200
[Credential-related source omitted] - Access to secretsline 201
[Credential-related source omitted] - Access to secretsline 202
[Credential-related source omitted] - Network accessline 219
对于 DeepSeek,设置 `VISUAL_RECAP_AGENT=openai-compatible`、 `VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`、 - Network accessline 220
`VISUAL_RECAP_BASE_URL=https://api.deepseek.com/v1`、 - Access to secretsline 221
[Credential-related source omitted] - Access to secretsline 240
[Credential-related source omitted] - Access to secretsline 258
[Credential-related source omitted] - Network accessline 269
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 285
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 297
[Credential-related source omitted] - Access to secretsline 298
[Credential-related source omitted] - Network accessline 304
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | 变量,仅适用于复制版本(可复用的调用方会忽略它)。设为 `false` 后,不再在每次推送时生成回顾。之后只在 PR 打开、重新打开、标记为可审阅或合并时,以及打上 `recap` 标签时生成回顾。仅打上必需标签不会触发回顾。 | | `RECAP_CLI_VERSION` - Network accessline 306
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | Secret。仅当把 Plans 应用自托管在不同源上时才需要。 - Network accessline 347
要安装它,把这个文件从 [BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Network accessline 348
[BuilderIO/agent-native](https://github.com/BuilderIO/agent-native/blob/main/.github/workflows/pr-visual-recap-fork.yml) - Access to secretsline 351
[Credential-related source omitted] - Access to secretsline 565
[Credential-related source omitted] - Access to secretsline 630
[Credential-related source omitted] - Access to secretsline 631
[Credential-related source omitted] - Access to secretsline 632
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/privacy-and-data-rights.mdx
- Network accessline 59
- [GDPR 正文](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [欧盟委员会:处理个人请求](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 60
- [欧盟委员会:处理个人请求](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [California Attorney General:CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [CCPA 条例](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/locales/zh-CN/real-time-collaboration.mdx
packages/core/docs/content/locales/zh-CN/recurring-jobs.mdx
packages/core/docs/content/locales/zh-CN/render.mdx
packages/core/docs/content/locales/zh-CN/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- 切勿使用 `innerHTML`、`eval()` 或 `document.write()` - Access to secretsline 391
[Credential-related source omitted] - Access to secretsline 407
[Credential-related source omitted] - Access to secretsline 413
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/server-database.mdx
- Network accessline 8
每个代理本机应用都把状态存储在 PostgreSQL 中。UI 和代理通过同一套 Actions、使用同一个 [Drizzle ORM](https://orm.drizzle.team) 客户端来读写同一批表。实时同步使用 SSE,并以轮询作为后备,因此无需手动刷新,变化就会显示在另一端。 - Network accessline 81
[PGlite](https://pglite.dev/) 在进程内运行一个基于 WASM 的 Postgres 数据库。它能让你在部署之前发现仅存在于 Postgres 的 schema 问题,同时保持和 PGlite 默认方案一样简单的搭建流程。与 PGlite 一样,它只是本地存储,不要用于生产或共享环境。
packages/core/docs/content/locales/zh-CN/server-middleware.mdx
- Network accessline 104
关于更多中间件模式——条件逻辑、异步操作、响应拦截——请参见 [Nitro 中间件文档](https://nitro.build/guide/routing#middleware)和 [H3 文档](https://h3.unjs.io)。 - Network accessline 147
- [**路由**](/docs/server-routes):受这个中间件保护的自定义路由 - [**安全性**](/docs/security):数据范围限定、访问守卫和完整的身份验证模型 - [**Nitro 中间件文档**](https://nitro.build/guide/routing#middleware):完整的中间件 API 与模式 - Network accessline 149
- [**Nitro 中间件文档**](https://nitro.build/guide/routing#middleware):完整的中间件 API 与模式
packages/core/docs/content/locales/zh-CN/server-overview.mdx
- Network accessline 204
[Nitro](https://nitro.build) 是驱动代理本机 HTTP 层的服务器工具集。你在 `server/` 中配置的大部分内容都属于 Nitro:基于文件的路由、启动插件和中间件。 - Network accessline 212
底层的请求/响应 API 是 [H3](https://h3.unjs.io)。当你在自定义路由中看到 `defineEventHandler`、`createError` 或 `getQuery` 时,那就是 H3。[Nitro 文档](https://nitro.build/guide)和 [H3 文档](https://h3.unjs.io)是框架未直接涵盖内容的参考资料。
packages/core/docs/content/locales/zh-CN/server-plugins.mdx
- Network accessline 137
插件可以使用 Nitro 的任何生命周期钩子:`request`、`beforeResponse`、`close` 等。[Nitro 插件文档](https://nitro.build/guide/plugins)详细介绍了完整的钩子 API 和插件约定。 - Network accessline 159
- [**路由**](/docs/server-routes):自定义文件路由及其使用时机 - [**编写 Agent 指令**](/docs/writing-agent-instructions):如何在 agent-chat 插件中撰写 `systemPrompt` - [**Nitro 插件文档**](https://nitro.build/guide/plugins):完整的插件 API 和生命周期钩子 - Network accessline 161
- [**Nitro 插件文档**](https://nitro.build/guide/plugins):完整的插件 API 和生命周期钩子
packages/core/docs/content/locales/zh-CN/server-routes.mdx
- Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/skills-guide.mdx
- Network accessline 315
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 316
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 317
"url": "https://assets.agent-native.com", - Network accessline 318
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/locales/zh-CN/template-analytics.mdx
packages/core/docs/content/locales/zh-CN/template-assets.mdx
- Network accessline 111
现场演示:[assets.agent-native.com](https://assets.agent-native.com)。 - Network accessline 460
<EmbeddedApp url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 461
url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 497
资产应用技能的应用 ID为 `assets` 并托管 MCP URL `https://assets.agent-native.com/mcp`. - Network accessline 498
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/locales/zh-CN/template-brain.mdx
- Network accessline 155
现场演示:[brain.agent-native.com](https://brain.agent-native.com)。
packages/core/docs/content/locales/zh-CN/template-calendar-developers.mdx
- Network accessline 26
这需要先安装 [Node.js](https://nodejs.org) 22.22 或更高版本,以及 - Network accessline 27
[pnpm](https://pnpm.io)。如果您还没有安装它们,或者还没有配置好 LLM 连 - Network accessline 37
应用会在 `http://localhost:8082` 打开。如果没有自动打开,请在浏览器中访 - Network accessline 46
1. **创建一个 Google Cloud 项目**,如果您还没有可用的项目:在 [Google Cloud Console](https://console.cloud.google.com/) 中,点击 - Network accessline 47
[Google Cloud Console](https://console.cloud.google.com/) 中,点击 - Network accessline 55
4. **创建凭证**:在 **APIs & Services → Credentials** 下,创建一个类型 为“Web application”的 OAuth 客户端 ID,并添加 `http://localhost:8082/_agent-native/google/callback` 作为授权重定 - Network accessline 56
为“Web application”的 OAuth 客户端 ID,并添加 `http://localhost:8082/_agent-native/google/callback` 作为授权重定 向 URI(8082 是日历默认的开发端口)。 - Network accessline 57
`http://localhost:8082/_agent-native/google/callback` 作为授权重定 - Access to secretsline 72
[Credential-related source omitted] - Network accessline 75
从 [Zoom OAuth 应用](https://marketplace.zoom.us/)(Zoom Marketplace →
packages/core/docs/content/locales/zh-CN/template-calendar-features.mdx
packages/core/docs/content/locales/zh-CN/template-calendar-integrations.mdx
- Access to secretsline 22
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/template-calendar.mdx
packages/core/docs/content/locales/zh-CN/template-chat-developers.mdx
- Network accessline 29
这需要先安装 [Node.js](https://nodejs.org) 22.22 或更高版本。如果 `pnpm`
packages/core/docs/content/locales/zh-CN/template-chat-integrations.mdx
- Access to secretsline 21
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/template-chat.mdx
packages/core/docs/content/locales/zh-CN/template-clips-developers.mdx
- Network accessline 22
这需要先安装 [Node.js](https://nodejs.org) 22.22 或更高版本,以及 [pnpm](https://pnpm.io)。如果你还没有安装它们中的任何一个,或者还没有配置好 LLM 连接(Builder.io、一个 Anthropic/OpenAI 密钥,或本地 Ollama),请参阅[快速入门](/docs/getting-started)。 - Network accessline 300
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 301
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 302
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/template-clips-embed.mdx
- Network accessline 40
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 53
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 54
[Credential-related source omitted] - Network accessline 56
const response = await fetch( - Network accessline 145
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 152
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/locales/zh-CN/template-clips-features.mdx
- Network accessline 117
桌面应用是为最流畅的日常捕获而构建的:屏幕和网络摄像头录制,带有一个可拖动的摄像头气泡,录制时进行实时转录,全程都不需要保持浏览器标签页打开。可以从 [clips.agent-native.com/download](https://clips.agent-native.com/download) 下载,支持 macOS、Windows 和 Linux。 - Network accessline 212
[Credential-related source omitted] - Network accessline 213
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/template-clips-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/template-clips.mdx
packages/core/docs/content/locales/zh-CN/template-content.mdx
- Network accessline 12
### [立即体验 →](https://content.agent-native.com)
packages/core/docs/content/locales/zh-CN/template-design-developers.mdx
packages/core/docs/content/locales/zh-CN/template-design-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted] - Network accessline 29
https://design.agent-native.com/mcp - Network accessline 32
Claude 和 Claude Desktop 会将其添加为一个自定义连接器,并通过登录来批准访问。Claude Code 添加同一个 URL,然后运行 `/mcp` 完成身份验证。Codex 以及其他 CLI 或 stdio 客户端则运行 `npx @agent-native/core@latest connect https://design.agent-native.com`,这会打开一个浏览器标签页来完成授权,并生成一个按用户区分、可随时撤销的令牌,不涉及任何共享密钥 - Network accessline 39
直重试。请在 Claude Code 中运行 `/mcp` 并重新连接,或者在任意终端中运 行 `npx -y @agent-native/core@latest reconnect https://design.agent-native.com`。两者都会重新完成身份验证,而无需重 - Network accessline 40
行 `npx -y @agent-native/core@latest reconnect https://design.agent-native.com`。两者都会重新完成身份验证,而无需重 新安装任何东西。 - Network accessline 41
https://design.agent-native.com`。两者都会重新完成身份验证,而无需重
packages/core/docs/content/locales/zh-CN/template-design.mdx
packages/core/docs/content/locales/zh-CN/template-dispatch.mdx
packages/core/docs/content/locales/zh-CN/template-forms-agent.mdx
- Network accessline 40
- **“发布这个,并把链接给我。”** 表单会立即上线,您会马上拿到它的公开 URL。 - **“创建一个匿名的客户反馈表单,并把公开链接给我。”** 表单从一开始就以匿名方式发布,因此永远不会记录提交者的信息。 - **“只允许从 https://example.com 提交这个表单。”** 表单开始拒绝来自其他任何地方的提交。目前这项功能只能通过请求代理来实现,构建器中还没有对应的开关。 - Network accessline 42
- **“只允许从 https://example.com 提交这个表单。”** 表单开始拒绝来自其他任何地方的提交。目前这项功能只能通过请求代理来实现,构建器中还没有对应的开关。
packages/core/docs/content/locales/zh-CN/template-forms-features.mdx
- Network accessline 14
[Credential-related source omitted] - Network accessline 15
[Credential-related source omitted] - Network accessline 94
公开提交可以要求先通过 [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/) 验证才能通过。在设置中填写 site key 和 secret key 即可开启此功能。即使不使用 Turnstile,也始终有两项防护在生效:一个会悄悄丢弃机器人提交的隐藏字段,以及一项会拒绝提交速度快于正常人填写表单所需时间的检查。 - Network accessline 110
目前构建器中还没有针对这项功能的开关。请让代理帮您设置,例如“只允许从 https://example.com 提交这个表单”。请为每个允许的网站提供准确的来源(协议加域名,例如 `https://example.com`,不带末尾路径)。 - Network accessline 116
[Credential-related source omitted] - Network accessline 117
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/template-forms.mdx
packages/core/docs/content/locales/zh-CN/template-mail.mdx
- Network accessline 97
现场演示:[mail.agent-native.com](https://mail.agent-native.com)。 - Network accessline 168
1. 打开[Google Cloud Console](https://console.cloud.google.com/)并创建一个项目。 - Network accessline 169
2. 在 APIs & Services → Library 下启用 **Gmail API**。 3. 创建 OAuth 2.0 凭证(类型:Web 应用程序)。添加 `http://localhost:8085/_agent-native/google/callback` 作为授权重定向 URI。 4. 将客户端 ID 和客户端密钥复制到正在运行的应用的“设置”页面,然后点击“**连接 Google 帐户**”。 - Network accessline 170
3. 创建 OAuth 2.0 凭证(类型:Web 应用程序)。添加 `http://localhost:8085/_agent-native/google/callback` 作为授权重定向 URI。
packages/core/docs/content/locales/zh-CN/template-plan.mdx
- Network accessline 130
这会安装三个 slash 命令——`/visual-plan`、`/visual-recap` 和 `/visualize-repo`——以及位于 `https://plan.agent-native.com/mcp` 的 - Network accessline 131
`/visualize-repo`——以及位于 `https://plan.agent-native.com/mcp` 的 - Network accessline 213
如果某个 Plans 工具返回了 `needs auth`、`Unauthorized` 或 `Session terminated`,不要不断重试。使用 `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 214
`Session terminated`,不要不断重试。使用 `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` 重新进行身份验证,或者对支持 OAuth 的主机重新运行 `/mcp` → - Network accessline 215
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`
packages/core/docs/content/locales/zh-CN/template-slides-features.mdx
packages/core/docs/content/locales/zh-CN/template-slides-integrations.mdx
- Access to secretsline 22
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/template-slides.mdx
packages/core/docs/content/locales/zh-CN/toolkit-capability-packages.mdx
packages/core/docs/content/locales/zh-CN/toolkit-settings.mdx
- Access to secretsline 122
[Credential-related source omitted] - Network accessline 190
此流程需要 OAuth 通过用户设备上的环回地址 `127.0.0.1` 接收 HTTP 回调,因此不适用于远程托管的应用;`localhost` 也不能替代 `127.0.0.1`。托管在 `*.agent-native.com` 上的网站无法使用此流程。需要合作伙伴批准的是托管应用的运营方,可通过 OpenAI 的[意向申请表](https://openai.com/form/sign-in-with-chatgpt-interest)申请。 - Network accessline 200
推理只能使用公开的 Responses API。ChatGPT 的私有后端端点和 ChatGPT 对话数据均不在此授权范围内。OpenAI 将令牌共享流程标记为预览版,因此可用模型、使用限制和支持的 API 功能都可能发生变化。请查看 OpenAI 的 [Sign in with ChatGPT 概览](https://developers.openai.com/siwc/token-sharing-open-source)、[注册和登录](https://developer
packages/core/docs/content/locales/zh-CN/tracking.mdx
- Network accessline 83
| 提供商 | 环境变量 | | -------- | ------------------------------------------------------------------------------------ | | PostHog | `POSTHOG_API_KEY`(必填)、`POSTHOG_HOS - Network accessline 84
| -------- | ------------------------------------------------------------------------------------ | | PostHog | `POSTHOG_API_KEY`(必填)、`POSTHOG_HOST`(可选,默认为`https://us.i.posthog.com`) | | 混合面板 | `MIXPANEL_TOKEN` - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Network accessline 205
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 206
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 207
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 225
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 226
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 283
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 284
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 285
fetch("https://analytics.example.com/events", { - Network accessline 343
- **批处理 HTTP** — 内置提供程序将事件排队并每 10 秒或累积 50 个事件时刷新(以先到者为准)。这可以最大限度地减少出站请求,而不会丢失数据。 - **无 SDK 依赖项** — 所有内置提供程序都使用原始 `fetch()`。没有 PostHog SDK,没有 Mixpanel SDK,没有 Amplitude SDK。保持框架轻量级。 - Network accessline 344
- **无 SDK 依赖项** — 所有内置提供程序都使用原始 `fetch()`。没有 PostHog SDK,没有 Mixpanel SDK,没有 Amplitude SDK。保持框架轻量级。
packages/core/docs/content/locales/zh-CN/turn-into-app.mdx
- Network accessline 12
- 一个能够运行终端命令的 AI 智能体,例如 Claude Code、Codex 或 Cursor。网页版 Claude 或 ChatGPT 可以改为把任务交给 Builder。请参阅 [运行位置](#hosts)。 - [Node.js](https://nodejs.org)。nodejs.org 的安装程序包含 `npx`,下方的安装命令由它运行。 - Network accessline 13
- [Node.js](https://nodejs.org)。nodejs.org 的安装程序包含 `npx`,下方的安装命令由它运行。 - Network accessline 18
- **在你的电脑上。** Node.js 22.22 或更高版本(用 `node --version` 检查),以及位于你的 `PATH` 中的 [pnpm](https://pnpm.io)。如果 `pnpm --version` 失败,请运行 `corepack enable`,它随 Node.js 一起提供。 - Network accessline 59
无论从哪种宿主出发,Builder 托管都需要 Agent-Native Dispatch MCP 连接器(`https://dispatch.agent-native.com/mcp`,OAuth),并调用其 `start-workspace-app-creation` 操作。你这边不会运行任何东西,因此在 Dispatch 返回的分支构建完成之前,报告会保持待处理或未验证状态。skill 只报告 Dispatch 实际返回的内容,绝不编造分支 URL。在没有该连接器的本地 - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/locales/zh-CN/updating-ui-in-production.mdx
- Network accessline 17
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment) 是一个独立的、具备代码能力的托管工作流,代理可以在其中于生产环境修改你应用的 UI。将 Builder.io 与你的代码仓库配合使用(提供免费套餐),然后直接提示进行 UI 更改。无需重新部署。
packages/core/docs/content/locales/zh-CN/upgrading-core-ui.mdx
- Access to secretsline 65
[Credential-related source omitted] - Access to secretsline 68
[Credential-related source omitted] - Network accessline 171
规范来源 URL: [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx)。 - Network accessline 172
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx)。
packages/core/docs/content/locales/zh-CN/vercel.mdx
- Access to secretsline 89
[Credential-related source omitted] - Network accessline 103
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 128
[Credential-related source omitted] - Network accessline 184
计划自动化、排队中的事件自动化以及被中断运行的清理,都在每分钟一次的清扫中完成。构建会添加一个 [Vercel cron 任务](https://vercel.com/docs/cron-jobs),通过 `/_agent-native/jobs/_process-sweep` 调用清扫。工作区构建会为每个应用各添加一个 cron 任务,路径为 `/<app-id>/_agent-native/jobs/_process-sweep`。
packages/core/docs/content/locales/zh-CN/voice-input.mdx
packages/core/docs/content/locales/zh-CN/what-is-agent-native.mdx
packages/core/docs/content/locales/zh-CN/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Network accessline 267
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 268
[Credential-related source omitted] - Network accessline 269
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 270
[Credential-related source omitted] - Access to secretsline 271
[Credential-related source omitted] - Access to secretsline 272
[Credential-related source omitted] - Network accessline 284
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 285
[Credential-related source omitted] - Network accessline 286
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 287
[Credential-related source omitted] - Access to secretsline 288
[Credential-related source omitted] - Access to secretsline 289
[Credential-related source omitted] - Network accessline 340
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 342
APP_URL=https://workspace.example.com
packages/core/docs/content/locales/zh-CN/workspace-management.mdx
- Network accessline 90
并不是每个需要进行更改的人都对 git 感到满意。 [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management) 支持可视化分支模型,可映射到底层的 git 分支 - 对于内容和副本更改、布局调整、设计迭代以及无需开发环境的 A/B 测试非常有用。 - Network accessline 140
关键提示:使用 GitHub 团队 (`@org/team`),而不是个人。框架和调度更改应始终需要平台审查。请参阅 [GitHub CODEOWNERS docs](https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/about-code-owners) 了解 glob 语法和多所有者模式。 - Network accessline 160
然后添加 [actions/labeler](https://github.com/actions/labeler) 操作 - 请参阅该存储库的 README 以了解完整的工作流程 YAML。打开或更新 PR 时,标签会自动应用。 - Access to secretsline 233
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/a2a-protocol.mdx
- Access to secretsline 93
[Credential-related source omitted] - Network accessline 117
"name": "Analytics Agent", "description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", - Network accessline 118
"description": "Runs analytics queries and returns chart data", "url": "https://analytics.example.com", "version": "1.0.0", - Network accessline 119
"url": "https://analytics.example.com", - Network accessline 314
const client = new A2AClient("https://analytics.example.com", "my-api-key"); - Network accessline 350
// One-shot: send text, get text back const response = await callAgent( "https://analytics.example.com", - Network accessline 351
const response = await callAgent( "https://analytics.example.com", "How many signups last week?", - Network accessline 352
"https://analytics.example.com", - Access to secretsline 354
[Credential-related source omitted] - Network accessline 370
"id": "foundry-research", "name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", - Network accessline 371
"name": "Foundry Research", "url": "https://example.services.ai.azure.com/a2a", "cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 372
"url": "https://example.services.ai.azure.com/a2a", - Network accessline 373
"cardUrl": "https://example.services.ai.azure.com/agentCard/v0.3", - Network accessline 408
[Azure AI Foundry 託管代理](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 413
Foundry 可以透過 [Agent Service](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint) - Network accessline 422
[Gemini Enterprise 可以註冊外部 A2A 代理](https://docs.cloud.google.com/gemini/enterprise/docs/register-and-manage-an-a2a-agent), - Network accessline 424
探索使用 `GET {url}/v1/card`,呼叫使用 `POST {url}/v1/message:send`,串流 使用 `POST {url}/v1/message:stream`。要求使用 Google OAuth 或 ADC bearer 權杖和 Google Cloud IAM 權限。端點和權限要求請參閱[使用註冊表 A2A 端點呼叫代理](https://docs.cloud.google.com/gemini/enterprise/docs/invoke- - Network accessline 426
權杖和 Google Cloud IAM 權限。端點和權限要求請參閱[使用註冊表 A2A 端點呼叫代理](https://docs.cloud.google.com/gemini/enterprise/docs/invoke-agent-a2a)。 - Network accessline 441
"id": "anthropic-research", "name": "Anthropic Research", "url": "https://api.anthropic.com", - Network accessline 442
"name": "Anthropic Research", "url": "https://api.anthropic.com", "kind": { - Network accessline 443
"url": "https://api.anthropic.com", - Access to secretsline 448
[Credential-related source omitted] - Access to secretsline 461
[Credential-related source omitted] - Network accessline 507
# 直接透過 URL 呼叫已部署的應用程式,從環境變數讀取不記名權杖 agent-native invoke https://analytics.example.com "How many signups last week?" \ - Network accessline 508
[Credential-related source omitted] - Access to secretsline 508
[Credential-related source omitted] - Access to secretsline 509
[Credential-related source omitted] - Access to secretsline 577
[Credential-related source omitted] - Access to secretsline 581
[Credential-related source omitted] - Network accessline 639
- 共用的 `A2A_SECRET` 在建置時安裝到每個應用程式的環境中。 - 跨應用呼叫是同來源的 - `https://workspace.example.com/apps/analytics` 呼叫 `https://workspace.example.com/apps/mail` - 因此沒有 DNS、CORS 或每對 JWT 設定。 - Network accessline 640
- 跨應用呼叫是同來源的 - `https://workspace.example.com/apps/analytics` 呼叫 `https://workspace.example.com/apps/mail` - 因此沒有 DNS、CORS 或每對 JWT 設定。 - Network accessline 648
**永遠不要依賴“一勞永逸”的 `Promise` 超過回應的壽命。**無伺服器函式(Netlify、Vercel、AWS Lambda、Cloud Run)會在回應內文重新整理時凍結 - 有時甚至在未等待的 `fetch(...)` 的 TCP 握手完成之前。在 Node 上本機工作的模式將默默地放棄正式環境中的工作。 - Network accessline 686
async run({ question }) { const response = await callAgent( "https://analytics.example.com", - Network accessline 687
const response = await callAgent( "https://analytics.example.com", question, - Network accessline 688
"https://analytics.example.com", - Access to secretsline 690
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/actions-defining.mdx
- Network accessline 221
`schema` 接受任何相容於 [Standard Schema](https://standardschema.dev) 的函式庫:
packages/core/docs/content/locales/zh-TW/add-ui-to-your-agent.mdx
packages/core/docs/content/locales/zh-TW/agent-mentions.mdx
packages/core/docs/content/locales/zh-TW/agent-teams.mdx
- Access to secretsline 138
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/agent-web-surfaces.mdx
- Network accessline 171
const files = buildAgentWebStaticFiles({ siteName: "My Agent-Native App", siteUrl: "https://example.com", - Network accessline 172
siteName: "My Agent-Native App", siteUrl: "https://example.com", description: "Public docs for my app.", - Network accessline 173
siteUrl: "https://example.com", - Network accessline 198
npx @agent-native/core@latest audit-agent-web --url https://www.agent-native.com
packages/core/docs/content/locales/zh-TW/audit-log.mdx
packages/core/docs/content/locales/zh-TW/authentication.mdx
- Network accessline 8
Agent-Native 應用程式使用 [Better Auth](https://better-auth.com) 進行帳戶優先設計的驗證。使用者在第一次存取時建立一個帳戶,並從第一天起就獲得真實身分。 - Access to secretsline 193
[Credential-related source omitted] - Access to secretsline 198
[Credential-related source omitted] - Access to secretsline 202
[Credential-related source omitted] - Access to secretsline 207
[Credential-related source omitted] - Access to secretsline 209
[Credential-related source omitted] - Access to secretsline 231
[Credential-related source omitted] - Access to secretsline 235
[Credential-related source omitted] - Network accessline 248
https://mail.agent-native.com/mcp - Access to secretsline 473
[Credential-related source omitted] - Access to secretsline 475
[Credential-related source omitted] - Access to secretsline 477
[Credential-related source omitted] - Access to secretsline 478
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/automations.mdx
packages/core/docs/content/locales/zh-TW/aws-amplify.mdx
- Access to secretsline 106
[Credential-related source omitted] - Network accessline 120
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 178
[Credential-related source omitted] - Network accessline 192
在 [Amplify 主控台](https://console.aws.amazon.com/amplify/)中建立應用程式: - Network accessline 226
第一次建置需要幾分鐘,你可以在 Amplify 主控台中追蹤進度。建置完成後,開啟應用程式 的網址,形式類似 `https://main.d1m7bkiki6tdw1.amplifyapp.com`,並建立第一個帳號。 - Network accessline 227
的網址,形式類似 `https://main.d1m7bkiki6tdw1.amplifyapp.com`,並建立第一個帳號。 - Access to secretsline 274
[Credential-related source omitted] - Access to secretsline 278
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/aws-lambda.mdx
- Network accessline 35
你還需要一個 AWS 帳號,並在你的電腦上安裝並登入 [AWS CLI](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html), - Access to secretsline 106
[Credential-related source omitted] - Access to secretsline 136
[Credential-related source omitted] - Access to secretsline 154
[Credential-related source omitted] - Access to secretsline 199
[Credential-related source omitted] - Access to secretsline 207
[Credential-related source omitted] - Network accessline 229
第一個指令會輸出 `FunctionUrl`,看起來像 `https://abc123.lambda-url.us-east-1.on.aws/`。 - Network accessline 247
在 [CloudFront 主控台](https://console.aws.amazon.com/cloudfront/v4/home)中建立 - Network accessline 261
在散發的 **Origins** 分頁中建立一個來源。在 **Origin domain** 中,貼上步驟 5 中 函式 URL 的主機名稱,不要包含 `https://` 或結尾的斜線。將 **Protocol** 設定為 - Network accessline 262
函式 URL 的主機名稱,不要包含 `https://` 或結尾的斜線。將 **Protocol** 設定為 - Network accessline 294
export BETTER_AUTH_URL='https://d111111abcdef8.cloudfront.net' - Access to secretsline 296
[Credential-related source omitted] - Access to secretsline 298
[Credential-related source omitted] - Access to secretsline 314
[Credential-related source omitted] - Dangerous shell operationline 316
rm -f function.zip && (cd .output/server && zip -qr ../../function.zip .)
packages/core/docs/content/locales/zh-TW/azure-static-web-apps.mdx
packages/core/docs/content/locales/zh-TW/blueprint-installer.mdx
packages/core/docs/content/locales/zh-TW/cli-adapters.mdx
- Access to secretsline 99
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/client-agent-chat.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 26
<Image src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500" - Network accessline 27
src="https://cdn.builder.io/api/v1/image/assets%2FYJIGb4i01jvw0SRdL5Bt%2F59e297f868244e39ba5ad9c62a32281e?width=500&height=500"
packages/core/docs/content/locales/zh-TW/client-data.mdx
- Network accessline 12
[Credential-related source omitted] - Network accessline 13
[Credential-related source omitted] - Network accessline 32
<div class="diagram-node">元件</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 33
<div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> <div class="diagram-node">fetch("/api/leads/:id")</div> <div class="fmcmp-arrow diagram-muted" aria-hidden="true">↓</div> - Network accessline 34
<div class="diagram-node">fetch("/api/leads/:id")</div> - Network accessline 125
optional: true, description: "來自 [useQuery options](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 的所有內容,除了 `queryKey` 和 `queryFn`,這兩者由此掛鉤自行設定。", - Network accessline 126
description: "來自 [useQuery options](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 的所有內容,除了 `queryKey` 和 `queryFn`,這兩者由此掛鉤自行設定。", }, - Network accessline 127
"來自 [useQuery options](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 的所有內容,除了 `queryKey` 和 `queryFn`,這兩者由此掛鉤自行設定。", - Network accessline 137
此掛鉤旨在用於唯讀 actions。它是 [useQuery()](https://tanstack.com/query/latest/docs/framework/react/reference/useQuery) 函式的簡單包裝。 - Network accessline 173
optional: true, description: "來自 [useMutation options](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 的所有內容,除了 `mutationFn`,此項由此掛鉤自行設定。", - Network accessline 174
description: "來自 [useMutation options](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 的所有內容,除了 `mutationFn`,此項由此掛鉤自行設定。", fields: [ - Network accessline 175
"來自 [useMutation options](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 的所有內容,除了 `mutationFn`,此項由此掛鉤自行設定。", - Network accessline 201
此掛鉤旨在用於變更資料的 actions,例如建立、更新或傳送某項內容。它是 [useMutation()](https://tanstack.com/query/latest/docs/framework/react/reference/useMutation) 函式的簡單包裝。
packages/core/docs/content/locales/zh-TW/cloneable-saas.mdx
- Network accessline 46
https://github.com/owner/repo - Network accessline 61
npx --yes @agent-native/core@latest create my-app --template community:owner/repo#v1.2.0 npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 62
npx --yes @agent-native/core@latest create my-app --template https://github.com/owner/repo#v1.2.0 - Network accessline 70
npx --yes @agent-native/core@latest create my-app --template 'community:owner/repo?app=analytics#v1.2.0' npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 71
npx --yes @agent-native/core@latest create my-app --template 'https://github.com/owner/repo?app=analytics#v1.2.0' - Network accessline 132
希望加入清單?請提交[社群模板](https://github.com/BuilderIO/agent-native/issues/new?template=community-template.yml), - Access to secretsline 284
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/cloudflare.mdx
- Access to secretsline 91
[Credential-related source omitted] - Network accessline 105
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 150
[Credential-related source omitted] - Network accessline 189
排程自動化、佇列中的事件自動化,以及被中斷執行的清理,都會在每分鐘一次的掃描中完成。建置會在產生的 `wrangler.json` 中新增一個 `* * * * *` [Cron Trigger](https://developers.cloudflare.com/workers/configuration/cron-triggers/),並保留你已經宣告的所有 Cron Trigger。觸發時,Worker 的 `scheduled` 處理常式會使用一個以 `A2A_SEC
packages/core/docs/content/locales/zh-TW/code-agents-ui.mdx
- Network accessline 202
npx @agent-native/core@latest code /migrate ./legacy-app --emit ./migration-dossier npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 203
npx @agent-native/core@latest code /audit --url https://example.com - Network accessline 355
npx @agent-native/core@latest code /migrate ./my-next-app --out ../migrated-app npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard" - Network accessline 356
npx @agent-native/core@latest code /migrate https://example.com --describe "marketing site plus dashboard"
packages/core/docs/content/locales/zh-TW/create-with-an-agent.mdx
- Network accessline 14
- 一個能執行終端機指令的 AI 代理,例如 Claude Code、Codex 或 Cursor。只能聊天的代理,例如 ChatGPT 或網頁版 Claude,也能提供協助。請參閱[無法執行指令的代理](#chat-only-agents)。 - [Node.js](https://nodejs.org) 22.22 或更新版本。 - Network accessline 15
- [Node.js](https://nodejs.org) 22.22 或更新版本。 - Network accessline 38
5. 用 `pnpm dev` 啟動應用程式,並讓它在背景持續執行。 6. 讀取 `pnpm dev` 輸出的本機 URL。檢查它是否有回應, 例如用 `curl -I <url>`。不要開啟瀏覽器,也不要使用無頭 - Network accessline 39
6. 讀取 `pnpm dev` 輸出的本機 URL。檢查它是否有回應, 例如用 `curl -I <url>`。不要開啟瀏覽器,也不要使用無頭 瀏覽器工具。我會自己開啟應用程式。 - Network accessline 40
例如用 `curl -I <url>`。不要開啟瀏覽器,也不要使用無頭 - Access to secretsline 49
[Credential-related source omitted] - Network accessline 57
- 一個新的應用程式資料夾,除非您改過名稱,否則叫 `my-app`,其中包含 `actions/`、`app/` 和一個 `AGENTS.md` 檔案。 - `pnpm dev` 正常執行且沒有錯誤。 - 一個本機 URL,通常是 `http://localhost:8080`,代理已經檢查過它有回應。 - Network accessline 59
- 一個本機 URL,通常是 `http://localhost:8080`,代理已經檢查過它有回應。
packages/core/docs/content/locales/zh-TW/creating-templates.mdx
packages/core/docs/content/locales/zh-TW/decisions.mdx
- Network accessline 18
{ type: "text", text: "Please cancel Friday's lunch reservation." }, { type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, - Network accessline 19
{ type: "json", value: { reservationId: "reservation-123" } }, { type: "image", url: "https://example.com/reservation.png" }, ]; - Network accessline 20
{ type: "image", url: "https://example.com/reservation.png" },
packages/core/docs/content/locales/zh-TW/deno-deploy.mdx
- Network accessline 24
你還需要一個 Deno Deploy 帳號,並在你的電腦上安裝 [Deno](https://docs.deno.com/runtime/getting_started/installation/)。 - Access to secretsline 93
[Credential-related source omitted] - Network accessline 107
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 154
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/deployment-environment-variables.mdx
- Network accessline 30
| `AGENT_NATIVE_SKIP_ENSURE_TABLES` | 在自動無伺服器跳過機制沒有涵蓋到的部署環境中,跳過隨選結構描述 DDL。這需要你的結構描述已經完成遷移。請參閱 [On-demand schema DDL](/docs/deployment#skip-ensure-tables)。 - Network accessline 32
| `AGENT_NATIVE_BUILDER_RELAY_TARGET_ORIGINS` | 僅限企業回呼使用、以逗號分隔的可信任預覽來源精確允許清單(例如 `https://0123456789abcdef01234567--content.netlify.app`)。只在企業回呼部署上設定它。Netlify 來源必須使用不可變、特定部署的 24 位十六進位永久連結。可變動的 `deploy-preview-*` 別名、萬用字元和網域字尾都會被拒絕。 | - Network accessline 45
| ------------------------------------------- | ------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------ - Network accessline 46
| `BETTER_AUTH_SECRET` | 32 個字元以上的隨機字串。用來簽署工作階段 cookies,也是 `OAUTH_STATE_SECRET` 和應用程式本機加密值的回退依據。它也是既有共用保管庫密文的舊版讀取候選值。硬性必要:如果在正式環境中缺少,框架會在啟動時擲出例外。 | | `BETTER_AUTH_URL` - Network accessline 47
| `BETTER_AUTH_URL` | 這個應用程式的選用公開來源覆寫值(例如 `https://mail.example.com`)。未設定時,框架會從請求、已知範本和託管平台的脈絡推斷出來。 - Access to secretsline 48
[Credential-related source omitted] - Access to secretsline 57
[Credential-related source omitted] - Access to secretsline 67
[Credential-related source omitted] - Access to secretsline 69
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/deployment.mdx
packages/core/docs/content/locales/zh-TW/dispatch.mdx
- Network accessline 116
Dispatch 可以是外部代理的單一 MCP 連線器:在 Claude、ChatGPT、Codex 或 Cursor 中新增一次 `https://dispatch.agent-native.com/mcp`,一次授權可到達每個授予的工作區應用程式,而不是每個應用程式一個連線器。請參閱 [External Agents](/docs/external-agents) 了解完整的連線流程、應用程式授權、OAuth 和行內 MCP 應用程式預覽。 - Network accessline 125
The single MCP connector URL external agents add (e.g. `https://dispatch.agent-native.com/mcp`). One authorization here reaches every **granted** workspace app instead of wiring one connector per app. App grants, OAuth, and inline MCP App p
packages/core/docs/content/locales/zh-TW/docker.mdx
- Dangerous shell operationline 66
RUN apt-get update && apt-get install -y --no-install-recommends python3 make g++ && rm -rf /var/lib/apt/lists/* - Dangerous shell operationline 68
&& rm -rf /var/lib/apt/lists/* - Access to secretsline 107
[Credential-related source omitted] - Access to secretsline 158
[Credential-related source omitted] - Network accessline 173
export BETTER_AUTH_URL='https://your-domain.example'
packages/core/docs/content/locales/zh-TW/durable-background-runs.mdx
- Network accessline 120
分派預設會送到部署自己的公開 URL(先是平台 URL,然後是 `APP_URL`),這需要繞經主機的邊緣網路一趟。將 `AGENT_NATIVE_SELF_DISPATCH_URL` 設為此程序監聽的位址,例如 `http://127.0.0.1:3000`,即可改走 loopback 分派。它必須是絕對的 `http(s)` URL,並會附加應用的 base path。它只用於部署送給自己的請求,絕不用於外部服務會呼叫的 URL。 - Network accessline 172
Netlify 背景函式適用於所有以點數計費的 Netlify 方案(包含 Free)以及 Enterprise。它們以函式運算量計費,與同步函式使用同一個計量:依函式記憶體大小(預設 1 GB),每 GB-hour 執行時間 10 點。Free 方案每月 300 點是硬性上限,點數用完後 Netlify 會暫停該團隊的專案。舊版方案(2025 年 9 月 4 日前建立的帳戶)依請求次數計費函式,且舊版 Free 與 Starter 方案不含背景函式;請參閱 Netlify
packages/core/docs/content/locales/zh-TW/embedding-sdk.mdx
- Network accessline 222
return ( <EmbeddedApp url="https://assets.agent-native.com/picker" - Network accessline 223
<EmbeddedApp url="https://assets.agent-native.com/picker" className="h-full w-full" - Network accessline 224
url="https://assets.agent-native.com/picker" - Network accessline 277
getMcpUrl("https://assets.agent-native.com"); - Network accessline 278
getA2AUrl("https://assets.agent-native.com"); - Network accessline 280
for await (const chunk of sendMessage( "https://assets.agent-native.com", - Network accessline 281
"https://assets.agent-native.com", - Network accessline 302
return ( <AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 303
<AgentNative agentUrl="https://agent.example.com/workspaces/acme/sidecar" className="h-full w-full" - Network accessline 304
agentUrl="https://agent.example.com/workspaces/acme/sidecar" - Network accessline 386
sendAgentNativeHostCommand, } from "@agent-native/core/client/host"; announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 388
announceAgentNativeFrameReady({ hostOrigin: "https://app.example.com" }); - Network accessline 390
const context = await requestAgentNativeHostContext({ hostOrigin: "https://app.example.com", - Network accessline 391
hostOrigin: "https://app.example.com", - Network accessline 394
const liveActions = await requestAgentNativeHostActions({ hostOrigin: "https://app.example.com", - Network accessline 395
hostOrigin: "https://app.example.com", - Network accessline 399
"select-element", { elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, - Network accessline 400
{ elementId: context.selection?.ids?.[0] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 401
{ hostOrigin: "https://app.example.com" }, - Network accessline 405
"refreshData", { queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, - Network accessline 406
{ queryKey: ["customer", context.resource?.id] }, { hostOrigin: "https://app.example.com" }, ); - Network accessline 407
{ hostOrigin: "https://app.example.com" }, - Network accessline 410
const hostTools = createAgentNativeHostTools({ hostOrigin: "https://app.example.com", - Network accessline 411
hostOrigin: "https://app.example.com", - Network accessline 464
useEffect(() => { const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", - Network accessline 465
const bridge = startAgentNativeBrowserSessionBridge({ hostOrigin: "https://app.example.com", label: "Builder editor", - Network accessline 466
hostOrigin: "https://app.example.com",
packages/core/docs/content/locales/zh-TW/extensions.mdx
- Network accessline 235
| `dbQuery(sql, args)` | 從 SQL 讀取(自動調整範圍給使用者) | `dbQuery('SELECT id, name FROM tools')` | | `dbExec(sql, args)` | Legacy trusted SQL writes for edi - Network accessline 236
| `dbExec(sql, args)` | Legacy trusted SQL writes for editor/admin extensions | Prefer `appAction` for app data writes | | `extensionFetch(url, options)` | 透過帶有秘密的安全代理攻擊外部 AP - Network accessline 237
| `extensionFetch(url, options)` | 透過帶有秘密的安全代理攻擊外部 API | `extensionFetch('https://api.github.com/user')` | - Network accessline 265
<script> const res = await extensionFetch('https://api.github.com/user', { - Network accessline 266
const res = await extensionFetch('https://api.github.com/user', {
packages/core/docs/content/locales/zh-TW/external-agents.mdx
- Network accessline 111
https://dispatch.agent-native.com/mcp - Network accessline 122
https://mail.agent-native.com/mcp - Network accessline 126
每個託管應用程式還有一個幫助頁面 `https://<app>/mcp/connect` 與可複製的 URL 和 - Network accessline 127
`https://<app>/mcp/connect` 與可複製的 URL 和 - Network accessline 214
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 258
原始 `skills` CLI 僅安裝 `SKILL.md` 檔案;本機 MCP 用戶端仍然 需要一個連線器,例如 `npx @agent-native/core@latest connect https://assets.agent-native.com`。 - Network accessline 259
需要一個連線器,例如 `npx @agent-native/core@latest connect https://assets.agent-native.com`。 - Network accessline 265
預設用戶端選取是所有支援的本機用戶端;新增 `--client codex`、`--client claude-code` 或其他特定目標以縮小設定範圍。行內主機(ChatGPT、網頁版、桌面版和行動版 Claude 聊天,以及 Claude Desktop 的 Code 分頁)會在聊天中顯示選取器或變體網格;CLI/僅連結主機(Codex CLI、Claude Code 終端機、Cowork、claude.ai 網頁版 Code)會傳回「在…中開啟 →」連結。使用者在瀏覽器 - Network accessline 271
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 286
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 289
`reconnect` 尋找給定主機和選定用戶端的 URL 以 `/mcp` 結尾的任何 MCP 設定條目(透過 URL 進行匹配,無論連線器名稱如何),然後重新整理或替換驗證材料,而無需觸及已安裝的 skills 或重新執行完整安裝流程。傳遞基本應用程式 URL(例如 `https://plan.agent-native.com`) - 推斷出 `/mcp` 後綴。驗證和工具載入是針對每個用戶端的,因此之後重新啟動/重新載入該用戶端; Codex 在新載入的工具出現之前需要 - Network accessline 297
對於無法直接新增遠端 OAuth URL 的 MCP 用戶端,請在瀏覽器中開啟應用程式並使用其 **Connect** 功能(在 `https://<app>/mcp/connect` 上提供)。登入後,點選“**連線/授權**”。該頁面為您提供一個設定偵測到的代理的一鍵深層連結,或一個可立即貼上的 `.mcp.json` 塊: - Network accessline 303
"mail": { "type": "http", "url": "https://mail.agent-native.com/mcp", - Network accessline 304
[Credential-related source omitted] - Network accessline 305
"url": "https://mail.agent-native.com/mcp", - Network accessline 321
claude mcp add --transport http agent-native \ https://dispatch.agent-native.com/mcp - Network accessline 322
https://dispatch.agent-native.com/mcp - Network accessline 325
這與 `npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code` 為您編寫的僅 URL 條目相同。然後在 Claude 程式碼中執行`/mcp`並選取**驗證**。用戶端從 MCP 伺服器的 `401 WWW-Authenticate` 質詢中發現驗證,取得 `/.well-known/oauth-protected-resource` 和 `/. - Network accessline 328
https://dispatch.agent-native.com/mcp - Network accessline 468
npx @agent-native/core@latest 連線 https://plan.agent-native.com --client codex --full-catalog - Network accessline 519
Claude Code calls: manage-draft(to: "john@example.com", subject: "Q3 Report", body: "…") → Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Network accessline 520
→ Open draft in Mail → https://mail.agent-native.com/_agent-native/open?app=mail&view=inbox&compose=… - Access to secretsline 739
[Credential-related source omitted] - Access to secretsline 755
[Credential-related source omitted] - Access to secretsline 761
[Credential-related source omitted] - Network accessline 767
"analytics": { "type": "http", "url": "https://analytics.agent-native.com/mcp", - Network accessline 768
[Credential-related source omitted] - Network accessline 769
"url": "https://analytics.agent-native.com/mcp", - Access to secretsline 770
[Credential-related source omitted] - Network accessline 792
`connect dev` 將相同的穩定 MCP 伺服器名稱(`agent-native-mail`、`agent-native-calendar` 等)重寫到本機 dev-lazy 閘道,因此工具名稱不會更改。在寫入開發條目之前,它會備份 `~/.agent-native/connect-profiles.json` 中的目前正式環境條目。預設閘道為`http://127.0.0.1:8080`;如果您的閘道行動了,請使用 `--gateway <url>` 或 `--po - Network accessline 819
- 使用 `npx @agent-native/core@latest connect https://dispatch.agent-native.com` 將您自己的代理連線到 Dispatch;僅當您需要一個獨立的應用程式時才使用直接應用程式 URL。 - Access to secretsline 839
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/file-uploads.mdx
- Network accessline 131
只有在需要使用不同種類的後端時,才需要註冊自訂提供者。以下範例會使用 `fetch()` 透過一般的 HTTP API 上傳檔案: - Network accessline 148
// and returns { url }. Adjust the request and response handling to // match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); - Network accessline 149
// match your own provider's actual contract. const response = await fetch(endpoint, { method: "POST", body }); if (!response.ok) { - Network accessline 150
const response = await fetch(endpoint, { method: "POST", body });
packages/core/docs/content/locales/zh-TW/getting-started.mdx
packages/core/docs/content/locales/zh-TW/harness-agents.mdx
- Network accessline 111
Agent-Native 可以充當[ACP](https://agentclientprotocol.com)(代理用戶端
packages/core/docs/content/locales/zh-TW/key-concepts.mdx
- Network accessline 328
}), run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); - Network accessline 329
run: async ({ source }) => { const res = await fetch(`https://api.example.com/${source}`); return await res.json(); - Network accessline 330
const res = await fetch(`https://api.example.com/${source}`);
packages/core/docs/content/locales/zh-TW/koyeb.mdx
packages/core/docs/content/locales/zh-TW/mcp-apps.mdx
- Network accessline 23
OpenAI 外掛擴充功能會在模型呼叫的 MCP 工具之外新增應用程式啟動點。啟用行內 MCP Apps 後,共用的 `open_app` 工具會同時提供全域側邊欄和對話側邊面板的進入點。ChatGPT 會以 `{}` 呼叫該工具;Core 會以行內 MCP App 開啟目前應用程式的首頁。使用共用 MCP 伺服器的應用程式不需要個別註冊。請參閱 [OpenAI 外掛擴充功能指南](https://developers.openai.com/plugins/build/ext - Network accessline 261
支援行內渲染 MCP Apps 的主機包括網頁版、桌面版和行動版 Claude 聊天,以及 Claude Desktop 的 [Code](https://github.com/anthropics/claude-code/issues/103809) 分頁; ChatGPT; VS Code GitHub Copilot (1.109+); Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex - Network accessline 263
在 Claude Code 終端機、Cowork、claude.ai 網頁版 Code 和 Codex CLI 中,請使用深層連結備援方案。[Cowork 尚未宣告支援 MCP Apps,在同一連接器上使用它可能會導致 Chat 中的行內應用程式消失,直到您重新連接該連接器。](https://github.com/anthropics/claude-code/issues/103924) 當主機選取不渲染 iframe 時,深層連結仍然是可靠的後備方案。實際上,每個 Age - Access to secretsline 269
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/mcp-clients.mdx
- Network accessline 148
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 149
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 178
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 179
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 180
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 226
[Credential-related source omitted] - Network accessline 269
Atlassian 預設使用 Atlassian Rovo MCP 的 OAuth 2.1 端點 `https://mcp.atlassian.com/v1/mcp/authv2`,並提供 Jira 和 Confluence - Network accessline 270
`https://mcp.atlassian.com/v1/mcp/authv2`,並提供 Jira 和 Confluence - Network accessline 276
Rovo MCP server → Domains** 下的受信任網域清單,並在 **Permissions** 中啟用 所需的 **Read**、**Write** 和 **Search** 權限。該網域必須是 Agent-Native 應用程式實際執行的主機;除非 `https://builder.io/**` 是部署的實際來源,否則 - Network accessline 277
所需的 **Read**、**Write** 和 **Search** 權限。該網域必須是 Agent-Native 應用程式實際執行的主機;除非 `https://builder.io/**` 是部署的實際來源,否則 不要將其複製到其他部署。請參閱 Atlassian 的 [Rovo MCP 設定指南](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 278
應用程式實際執行的主機;除非 `https://builder.io/**` 是部署的實際來源,否則 - Network accessline 279
不要將其複製到其他部署。請參閱 Atlassian 的 [Rovo MCP 設定指南](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 280
和[管理員控制說明](https://support.atlassian.com/security-and-access-policies/docs/control-atlassian-rovo-mcp-server-settings/)。 - Network accessline 289
1. **伺服器名稱** - 供您自己參考的簡短標籤(例如“zapier”、“slack-tools”)。 2. **URL** — MCP 伺服器提供者為您提供的 HTTPS 端點(例如 `https://mcp.zapier.com/s/abc123/mcp`)。這通常可以在提供者的儀表板或整合檔案中找到。 - Network accessline 290
2. **URL** — MCP 伺服器提供者為您提供的 HTTPS 端點(例如 `https://mcp.zapier.com/s/abc123/mcp`)。這通常可以在提供者的儀表板或整合檔案中找到。 - Network accessline 305
[Credential-related source omitted] - Network accessline 397
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 398
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 399
"url": "https://example.com/mcp", - Network accessline 437
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com
packages/core/docs/content/locales/zh-TW/mcp-protocol.mdx
- Access to secretsline 31
[Credential-related source omitted] - Network accessline 108
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 117
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 118
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 119
"url": "https://mail.example.com/mcp", - Network accessline 138
POST https://your-app.example.com/mcp - Access to secretsline 143
[Credential-related source omitted] - Access to secretsline 254
[Credential-related source omitted] - Network accessline 262
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 268
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 269
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 384
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 393
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 394
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 395
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/locales/zh-TW/messaging.mdx
- Network accessline 81
1. 轉到 **[api.slack.com/apps](https://api.slack.com/apps)** 並點選 **建立新應用程式** → **從頭開始**。為其命名(例如“Agent”)並選取您的工作區。 - Network accessline 97
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 120
https://your-clips.example.com/api/slack/unfurl - Access to secretsline 134
[Credential-related source omitted] - Network accessline 166
將 OAuth Redirect URL 設為 `https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 167
`https://your-app.example.com/_agent-native/integrations/slack/oauth/callback`, - Network accessline 168
Event Subscriptions Request URL 設為 `https://your-app.example.com/_agent-native/integrations/slack/webhook`, Interactivity Request URL 設為 - Network accessline 169
`https://your-app.example.com/_agent-native/integrations/slack/webhook`, - Network accessline 170
Interactivity Request URL 設為 `https://your-app.example.com/_agent-native/integrations/slack/interactions`。 訂閱 `app_home_opened`、`app_context_changed`、`app_mention` 和 `message.im`。 - Network accessline 171
`https://your-app.example.com/_agent-native/integrations/slack/interactions`。 - Access to secretsline 173
[Credential-related source omitted] - Network accessline 259
[Credential-related source omitted] - Access to secretsline 259
[Credential-related source omitted] - Access to secretsline 264
[Credential-related source omitted] - Access to secretsline 361
[Credential-related source omitted] - Network accessline 375
1. 開啟電報和訊息**[@BotFather](https://t.me/BotFather)**。 - Network accessline 382
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 401
1. 在 **[resend.com](https://resend.com)** 註冊。免費套餐足以開始使用。 - Network accessline 408
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 415
[Credential-related source omitted] - Network accessline 422
1. 在 **[sendgrid.com](https://sendgrid.com)** 註冊。 - Network accessline 430
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 435
[Credential-related source omitted] - Network accessline 459
1. 轉到 **[Meta Developer Portal](https://developers.facebook.com/)**,點選 **建立應用**,然後選取 **商業** 型別。 - Access to secretsline 466
[Credential-related source omitted] - Network accessline 472
回調 URL:https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Access to secretsline 683
[Credential-related source omitted] - Access to secretsline 684
[Credential-related source omitted] - Access to secretsline 794
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/multi-app-workspace.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 100
[Credential-related source omitted] - Network accessline 245
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 246
"urls": { "design": "https://design.example.test" } - Network accessline 247
"design": "https://design.example.test" - Access to secretsline 276
[Credential-related source omitted] - Access to secretsline 282
[Credential-related source omitted] - Access to secretsline 285
[Credential-related source omitted] - Network accessline 325
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 326
# https://your-agents.com/mail/* → apps/mail - Network accessline 327
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 328
# https://your-agents.com/forms/* → apps/forms
packages/core/docs/content/locales/zh-TW/native-chat-ui.mdx
- Possible obfuscationline 152
僅當操作聲明 `chatUI`**或**結果攜帶顯式已知的 `widget` 判別式時,渲染器才會接管。它從不形<U+200B><U+200B>狀推斷任意物件,也從不從工具結果執行 HTML 或 JavaScript - 因此本機小工具不能成為注入向量。 - Network accessline 446
舊 API 與新 API 的對照表及移轉步驟,請參閱 [AgentKit 聊天移轉指南](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md)。
packages/core/docs/content/locales/zh-TW/netlify.mdx
packages/core/docs/content/locales/zh-TW/node-js.mdx
packages/core/docs/content/locales/zh-TW/notifications.mdx
- Network accessline 192
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 193
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 194
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 211
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 212
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 213
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/locales/zh-TW/observability.mdx
- Network accessline 507
儲存庫中的 [`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel) 提供這個外掛程式更完整的版本,支援依訊號設定端點並驗證匯出器。它沒有發布到 npm,因此請從上面的外掛程式開始,而不是安裝它。
packages/core/docs/content/locales/zh-TW/onboarding.mdx
- Access to secretsline 186
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/plan-plugin.mdx
- Network accessline 14
- **三個技能** — `/visual-plan`(規範入口點)、`/visual-recap` 和 `/visualize-repo`。 - **Plan MCP 連線器** — 針對 `https://plan.agent-native.com` 上的託管應用程式完成註冊(MCP 端點為 `https://plan.agent-native.com/mcp`,伺服器名稱為 `plan`)。 - Network accessline 15
- **Plan MCP 連線器** — 針對 `https://plan.agent-native.com` 上的託管應用程式完成註冊(MCP 端點為 `https://plan.agent-native.com/mcp`,伺服器名稱為 `plan`)。 - Network accessline 81
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 96
- **使用 Chromium 系瀏覽器。** Safari 會封鎖託管的 HTTPS Plan 頁面讀取 `http://127.0.0.1` 這個 localhost 橋接(混合內容/私人網路),導致頁面卡在「Loading plan.」。在 macOS 上,`--open` 已經會優先選用 Chrome/Chromium/Edge/Brave;如果 Safari 還是開啟了,請在 Chromium 系瀏覽器中重新開啟印出的網址。 - Network accessline 164
- `--client codex|claude-code|claude-code-cli|cowork|all` — 要為哪些本機代理寫入 MCP 設定(預設為 `all`)。 - `--no-connect` — 只註冊連線器而不進行驗證;之後再執行 `npx @agent-native/core@latest connect https://plan.agent-native.com --client all`,或選擇更精確的 `--client`。 - Network accessline 165
- `--no-connect` — 只註冊連線器而不進行驗證;之後再執行 `npx @agent-native/core@latest connect https://plan.agent-native.com --client all`,或選擇更精確的 `--client`。 - Network accessline 206
安裝後,**開啟一個新的 Codex 討論串**,讓技能與 MCP 工具載入到工作階段中。這個外掛提供一個僅含網址的連線器(`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`);`codex mcp login plan` 會執行 OAuth 流程。如果你偏好用一個指令同時完成安裝與驗證,或是想要本機檔案或自行託管模式,上面的通用 CLI 路徑(`npx @agent-native/core@latest sk - Network accessline 210
**舊版安裝:** 如果你的設定中還有一個指向同一個網址的 `agent-native-plans` 項目,對 Codex 執行 `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`,或用相同指令搭配你的目標 `--client`,就能把它整併成規範的 `plan` 名稱。
packages/core/docs/content/locales/zh-TW/pr-visual-recap.mdx
- Access to secretsline 140
[Credential-related source omitted] - Access to secretsline 141
[Credential-related source omitted] - Access to secretsline 142
[Credential-related source omitted] - Network accessline 157
[Credential-related source omitted] - Access to secretsline 157
[Credential-related source omitted] - Access to secretsline 172
[Credential-related source omitted] - Access to secretsline 187
[Credential-related source omitted] - Network accessline 192
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 202
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 212
[Credential-related source omitted] - Access to secretsline 213
[Credential-related source omitted] - Network accessline 219
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | 變數,僅適用於複製版本(可重複使用的呼叫端會忽略它)。設為 `false` 後,不再於每次推送時產生回顧。之後只會在 PR 開啟、重新開啟、標記為可審閱或合併時,以及套用 `recap` 標籤時產生回顧。僅套用必要標籤不會觸發回顧。 | | `RECAP_CLI_VERSION` - Network accessline 221
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | Secret。只有在把 Plan 應用程式自行託管於不同來源時才需要。 - Network accessline 241
[Credential-related source omitted] - Access to secretsline 241
[Credential-related source omitted] - Access to secretsline 414
[Credential-related source omitted] - Access to secretsline 469
[Credential-related source omitted] - Access to secretsline 470
[Credential-related source omitted] - Access to secretsline 471
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/privacy-and-data-rights.mdx
- Network accessline 59
- [GDPR 內容](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [歐盟執委會:處理個人要求](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 60
- [歐盟執委會:處理個人要求](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [California Attorney General:CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [CCPA 規則](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/locales/zh-TW/real-time-collaboration.mdx
packages/core/docs/content/locales/zh-TW/recurring-jobs.mdx
packages/core/docs/content/locales/zh-TW/render.mdx
packages/core/docs/content/locales/zh-TW/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- 切勿使用 `innerHTML`、`eval()` 或 `document.write()` - Access to secretsline 393
[Credential-related source omitted] - Access to secretsline 409
[Credential-related source omitted] - Access to secretsline 415
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/server-database.mdx
- Network accessline 8
每個 Agent-Native 應用都將其狀態儲存在 PostgreSQL 中。UI 與 agent 都透過同一組 actions,使用同一個 [Drizzle ORM](https://orm.drizzle.team) 用戶端,讀寫相同的資料表。即時同步使用 SSE,並以輪詢作為備援,因此不需手動重新整理,變更就會顯示在另一端。 - Network accessline 81
[PGlite](https://pglite.dev/) 在行程內執行一個 WASM 版本的 Postgres 資料庫。它讓您能在部署前就抓到僅限 Postgres 的架構問題,同時維持與 PGlite 預設值一樣簡單的設定方式。和 PGlite 一樣,它僅是本機儲存,請勿用於正式環境或共用環境。
packages/core/docs/content/locales/zh-TW/server-middleware.mdx
- Network accessline 104
若需要更多中間件模式——條件邏輯、非同步操作、回應攔截——請參閱 [Nitro 中間件文件](https://nitro.build/guide/routing#middleware) 與 [H3 文件](https://h3.unjs.io)。 - Network accessline 147
- [**路由**](/docs/server-routes):受此中間件保護的自訂路由 - [**Security**](/docs/security):資料範圍限定、存取守衛,以及完整的驗證模型 - [**Nitro 中間件文件**](https://nitro.build/guide/routing#middleware):完整的中間件 API 與模式 - Network accessline 149
- [**Nitro 中間件文件**](https://nitro.build/guide/routing#middleware):完整的中間件 API 與模式
packages/core/docs/content/locales/zh-TW/server-overview.mdx
- Network accessline 204
[Nitro](https://nitro.build) 是驅動 Agent-Native HTTP 層的伺服器工具套件。您在 `server/` 中設定的內容大多是 Nitro:基於檔案的路由、啟動外掛,以及中間件。 - Network accessline 212
底層的請求/回應 API 是 [H3](https://h3.unjs.io)。當您在自訂路由中看到 `defineEventHandler`、`createError` 或 `getQuery` 時,那就是 H3。[Nitro 文件](https://nitro.build/guide) 和 [H3 文件](https://h3.unjs.io) 是框架未直接涵蓋內容的參考來源。
packages/core/docs/content/locales/zh-TW/server-plugins.mdx
- Network accessline 139
外掛可以使用 Nitro 的任何生命週期掛鉤:`request`、`beforeResponse`、`close` 等等。[Nitro 外掛文件](https://nitro.build/guide/plugins) 詳細涵蓋了完整的掛鉤 API 與外掛約定。 - Network accessline 161
- [**路由**](/docs/server-routes):自訂檔案路由與使用時機 - [**Writing Agent Instructions**](/docs/writing-agent-instructions):如何在 agent-chat 外掛中撰寫 `systemPrompt` - [**Nitro 外掛文件**](https://nitro.build/guide/plugins):完整的外掛 API 與生命週期掛鉤 - Network accessline 163
- [**Nitro 外掛文件**](https://nitro.build/guide/plugins):完整的外掛 API 與生命週期掛鉤
packages/core/docs/content/locales/zh-TW/server-routes.mdx
- Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/skills-guide.mdx
- Network accessline 315
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 316
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 317
"url": "https://assets.agent-native.com", - Network accessline 318
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/locales/zh-TW/template-analytics.mdx
packages/core/docs/content/locales/zh-TW/template-assets.mdx
- Network accessline 111
線上示範:[assets.agent-native.com](https://assets.agent-native.com)。 - Network accessline 460
<EmbeddedApp url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 461
url="https://assets.agent-native.com/picker?mediaType=image" - Network accessline 497
資產應用技能的應用 ID為 `assets` 並託管 MCP URL `https://assets.agent-native.com/mcp`. - Network accessline 498
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/locales/zh-TW/template-brain.mdx
- Network accessline 155
線上示範:[brain.agent-native.com](https://brain.agent-native.com)。
packages/core/docs/content/locales/zh-TW/template-calendar-developers.mdx
- Network accessline 26
這需要先安裝 [Node.js](https://nodejs.org) 22.22 或更新版本,以及 - Network accessline 27
[pnpm](https://pnpm.io)。如果您還沒有安裝,或還沒有設定 LLM 連線 - Network accessline 37
應用程式會在 `http://localhost:8082` 開啟。如果沒有自動開啟,請在瀏覽器中 - Network accessline 46
1. **建立 Google Cloud 專案**(如果您還沒有可用的專案):在 [Google Cloud Console](https://console.cloud.google.com/) 中,點選頁面 - Network accessline 47
[Google Cloud Console](https://console.cloud.google.com/) 中,點選頁面 - Network accessline 55
4. **建立憑證**:在**APIs & Services → Credentials**下,建立一個型別為 「Web application」的 OAuth 用戶端 ID,並將 `http://localhost:8082/_agent-native/google/callback` 新增為已授權的 - Network accessline 56
「Web application」的 OAuth 用戶端 ID,並將 `http://localhost:8082/_agent-native/google/callback` 新增為已授權的 重新導向 URI(8082 是行事曆的預設開發連接埠)。 - Network accessline 57
`http://localhost:8082/_agent-native/google/callback` 新增為已授權的 - Access to secretsline 72
[Credential-related source omitted] - Network accessline 75
從一個 [Zoom OAuth 應用程式](https://marketplace.zoom.us/)取得用戶端 ID 和
packages/core/docs/content/locales/zh-TW/template-calendar-features.mdx
packages/core/docs/content/locales/zh-TW/template-calendar-integrations.mdx
- Access to secretsline 22
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/template-calendar.mdx
packages/core/docs/content/locales/zh-TW/template-chat-developers.mdx
- Network accessline 29
這需要先安裝 [Node.js](https://nodejs.org) 22.22 或更新版本。如果 `pnpm` 不在
packages/core/docs/content/locales/zh-TW/template-chat-integrations.mdx
- Access to secretsline 21
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/template-chat.mdx
packages/core/docs/content/locales/zh-TW/template-clips-developers.mdx
- Network accessline 22
這需要先安裝 [Node.js](https://nodejs.org) 22.22 或更新版本,以及 [pnpm](https://pnpm.io)。如果您還沒有安裝其中任何一項,或還沒有設定 LLM 連線(Builder.io、Anthropic/OpenAI 金鑰,或本機 Ollama),請參閱 [開始使用](/docs/getting-started)。 - Network accessline 300
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 301
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 302
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/template-clips-embed.mdx
- Network accessline 40
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 53
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 54
[Credential-related source omitted] - Network accessline 56
const response = await fetch( - Network accessline 145
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 152
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/locales/zh-TW/template-clips-features.mdx
- Network accessline 119
桌面應用程式的設計目標,是讓您每天的捕捉工作盡可能順暢:螢幕和網路攝影機錄製,搭配可拖曳的攝影機圓形視窗,並在您錄製時即時轉錄,全程不需要保持瀏覽器分頁開啟。可從 [clips.agent-native.com/download](https://clips.agent-native.com/download) 下載,支援 macOS、Windows 和 Linux。 - Network accessline 214
[Credential-related source omitted] - Network accessline 215
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/template-clips-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/template-clips.mdx
packages/core/docs/content/locales/zh-TW/template-content.mdx
- Network accessline 12
### [立即試用 →](https://content.agent-native.com)
packages/core/docs/content/locales/zh-TW/template-design-developers.mdx
packages/core/docs/content/locales/zh-TW/template-design-integrations.mdx
- Access to secretsline 22
[Credential-related source omitted] - Network accessline 43
https://design.agent-native.com/mcp - Network accessline 47
取權限。Claude Code 加入同一個網址後,執行 `/mcp` 進行驗證。Codex 和其 他 CLI 或 stdio 用戶端則改為執行 `npx @agent-native/core@latest connect https://design.agent-native.com`,這會開啟一個瀏覽器分頁進行授權,並產 - Network accessline 48
他 CLI 或 stdio 用戶端則改為執行 `npx @agent-native/core@latest connect https://design.agent-native.com`,這會開啟一個瀏覽器分頁進行授權,並產 生一個可撤銷的個人專屬權杖,不涉及任何共用密鑰。 - Network accessline 49
https://design.agent-native.com`,這會開啟一個瀏覽器分頁進行授權,並產 - Network accessline 61
重試。在 Claude Code 中執行 `/mcp` 並重新連線,或從任一終端機執行 `npx -y @agent-native/core@latest reconnect https://design.agent-native.com`。這兩種方式都能重新驗證,不需要重新 - Network accessline 62
`npx -y @agent-native/core@latest reconnect https://design.agent-native.com`。這兩種方式都能重新驗證,不需要重新 安裝任何東西。 - Network accessline 63
https://design.agent-native.com`。這兩種方式都能重新驗證,不需要重新
packages/core/docs/content/locales/zh-TW/template-design.mdx
packages/core/docs/content/locales/zh-TW/template-dispatch.mdx
packages/core/docs/content/locales/zh-TW/template-forms-agent.mdx
- Network accessline 40
- **「發布這份表單,並給我連結。」** 表單會立即上線,您也會馬上拿到它的公開網址。 - **「建立一個匿名的客戶意見回饋表單,並給我公開連結。」** 表單從一開始就以匿名方式發布,永遠不會擷取任何提交者資訊。 - **「只允許這份表單從 https://example.com 提交。」** 表單會開始拒絕來自其他任何地方的提交。目前這項功能只能透過請代理設定來使用,建置器中沒有對應的開關。 - Network accessline 42
- **「只允許這份表單從 https://example.com 提交。」** 表單會開始拒絕來自其他任何地方的提交。目前這項功能只能透過請代理設定來使用,建置器中沒有對應的開關。
packages/core/docs/content/locales/zh-TW/template-forms-features.mdx
- Network accessline 14
[Credential-related source omitted] - Network accessline 15
[Credential-related source omitted] - Network accessline 94
公開提交內容可以要求先通過 [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/) 驗證才能送出。在設定中設定網站金鑰和密鑰即可開啟這項功能。即使沒有設定 Turnstile,仍有兩項防護會持續運作:一個會悄悄捨棄機器人提交內容的隱藏欄位,以及一項會拒絕任何送出速度快到不像是真人合理填完表單的檢查。 - Network accessline 110
目前建置器中沒有這個功能的開關。請代理幫您設定,例如「只允許這份表單從 https://example.com 提交」。針對每個應該被允許的網站,請提供代理確切的來源(協定加網域,例如 `https://example.com`,不含結尾的路徑)。 - Network accessline 116
[Credential-related source omitted] - Network accessline 117
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/template-forms.mdx
packages/core/docs/content/locales/zh-TW/template-mail.mdx
- Network accessline 97
線上示範:[mail.agent-native.com](https://mail.agent-native.com)。 - Network accessline 168
1. 開啟[Google Cloud Console](https://console.cloud.google.com/)並建立一個專案。 - Network accessline 169
2. 在 APIs & Services → Library 下啟用 **Gmail API**。 3. 建立 OAuth 2.0 憑證(型別:Web 應用程式)。新增 `http://localhost:8085/_agent-native/google/callback` 作為授權重新導向 URI。 4. 將用戶端 ID 和用戶端金鑰複製到正在執行的應用的“設定”頁面,然後點選“**連線 Google 帳戶**”。 - Network accessline 170
3. 建立 OAuth 2.0 憑證(型別:Web 應用程式)。新增 `http://localhost:8085/_agent-native/google/callback` 作為授權重新導向 URI。
packages/core/docs/content/locales/zh-TW/template-plan.mdx
- Network accessline 113
這會安裝三個斜線指令 — `/visual-plan`、`/visual-recap` 和 `/visualize-repo` — 以及 `plan` MCP 連線器(代理與之溝通的遠端工具伺服器),位址為 `https://plan.agent-native.com/mcp`。完整的旗標參考、Claude Code 與 Codex 的外掛/市場安裝路徑,以及 VS Code 擴充功能,請見[Plan 外掛與市場 → 安裝路徑](/docs/plan-plugin#insta - Network accessline 163
如果 Plans 的工具傳回 `needs auth`、`Unauthorized` 或 `Session terminated`,不要一直重試。請用 `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` 重新驗證,或者對支援 OAuth 的主機重新執行 `/mcp` → **Authenticate**。之後請開啟一個新的 Codex 討論串,或重新
packages/core/docs/content/locales/zh-TW/template-slides-features.mdx
packages/core/docs/content/locales/zh-TW/template-slides-integrations.mdx
- Access to secretsline 22
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/template-slides.mdx
packages/core/docs/content/locales/zh-TW/toolkit-capability-packages.mdx
packages/core/docs/content/locales/zh-TW/toolkit-settings.mdx
- Access to secretsline 122
[Credential-related source omitted] - Network accessline 190
此流程需要透過使用者裝置上的迴路位址 `127.0.0.1` 接收 HTTP OAuth 回呼,因此不適用於遠端託管的應用程式;`localhost` 不能取代 `127.0.0.1`。託管於 `*.agent-native.com` 的網站無法使用此流程。需要申請合作夥伴核准的是託管應用程式的營運者,可透過 OpenAI 的[意願表單](https://openai.com/form/sign-in-with-chatgpt-interest)提出申請。 - Network accessline 200
推論只能使用公開的 Responses API。ChatGPT 私有後端端點和 ChatGPT 對話資料不在此授權範圍內。OpenAI 將權杖分享流程標示為預覽版,因此可用模型、限制和支援的 API 功能都可能變更。請參閱 OpenAI 的[使用 ChatGPT 登入概覽](https://developers.openai.com/siwc/token-sharing-open-source)、[註冊與登入](https://developers.openai.com/si
packages/core/docs/content/locales/zh-TW/tracking.mdx
- Network accessline 83
| 提供者 | 環境變數 | | -------- | ------------------------------------------------------------------------------------ | | PostHog | `POSTHOG_API_KEY`(必填)、`POSTHOG_HOS - Network accessline 84
| -------- | ------------------------------------------------------------------------------------ | | PostHog | `POSTHOG_API_KEY`(必填)、`POSTHOG_HOST`(可選,預設為`https://us.i.posthog.com`) | | 混合面板 | `MIXPANEL_TOKEN` - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Network accessline 205
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 206
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 207
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 225
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 226
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 283
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 284
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 285
fetch("https://analytics.example.com/events", { - Network accessline 343
- **批處理 HTTP** — 內建提供者將事件排隊並每 10 秒或累積 50 個事件時重新整理(以先到者為準)。這可以最大限度地減少出站請求,而不會丟失資料。 - **無 SDK 依賴項** — 所有內建提供者都使用原始 `fetch()`。沒有 PostHog SDK,沒有 Mixpanel SDK,沒有 Amplitude SDK。保持框架輕量級。 - Network accessline 344
- **無 SDK 依賴項** — 所有內建提供者都使用原始 `fetch()`。沒有 PostHog SDK,沒有 Mixpanel SDK,沒有 Amplitude SDK。保持框架輕量級。
packages/core/docs/content/locales/zh-TW/turn-into-app.mdx
- Network accessline 12
- 一個能夠執行終端機指令的 AI 代理,例如 Claude Code、Codex 或 Cursor。網頁版 Claude 或 ChatGPT 可以改為把工作交給 Builder。請參閱 [執行位置](#hosts)。 - [Node.js](https://nodejs.org)。nodejs.org 的安裝程式包含 `npx`,下方的安裝指令由它執行。 - Network accessline 13
- [Node.js](https://nodejs.org)。nodejs.org 的安裝程式包含 `npx`,下方的安裝指令由它執行。 - Network accessline 18
- **在你的電腦上。** Node.js 22.22 或更新版本(用 `node --version` 檢查),以及位於你的 `PATH` 中的 [pnpm](https://pnpm.io)。如果 `pnpm --version` 失敗,請執行 `corepack enable`,它隨 Node.js 一併提供。 - Network accessline 59
無論從哪種主機出發,Builder 託管都需要 Agent-Native Dispatch MCP 連接器(`https://dispatch.agent-native.com/mcp`,OAuth),並呼叫其 `start-workspace-app-creation` 動作。你這端不會執行任何東西,因此在 Dispatch 回傳的分支建構完成之前,報告會維持待處理或未驗證狀態。skill 只會回報 Dispatch 實際回傳的內容,絕不捏造分支 URL。在沒有該連接器的本 - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/locales/zh-TW/updating-ui-in-production.mdx
- Network accessline 17
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment) 是一個獨立的受管理、具備程式碼編輯能力的工作流程,讓代理可以在正式環境中修改你應用程式的 UI。將 Builder.io 與你的儲存庫搭配使用(提供免費方案),直接提示進行 UI 變更即可,不需要重新部署。
packages/core/docs/content/locales/zh-TW/upgrading-core-ui.mdx
- Access to secretsline 66
[Credential-related source omitted] - Access to secretsline 69
[Credential-related source omitted] - Network accessline 173
正式來源 URL: [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx)。 - Network accessline 174
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx)。
packages/core/docs/content/locales/zh-TW/vercel.mdx
- Access to secretsline 89
[Credential-related source omitted] - Network accessline 103
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 128
[Credential-related source omitted] - Network accessline 184
排程自動化、佇列中的事件自動化,以及被中斷執行的清理,都會在每分鐘一次的掃描中完成。建置會新增一個 [Vercel cron 工作](https://vercel.com/docs/cron-jobs),透過 `/_agent-native/jobs/_process-sweep` 呼叫掃描。工作區建置會為每個應用程式各新增一個 cron 工作,路徑為 `/<app-id>/_agent-native/jobs/_process-sweep`。
packages/core/docs/content/locales/zh-TW/voice-input.mdx
packages/core/docs/content/locales/zh-TW/what-is-agent-native.mdx
packages/core/docs/content/locales/zh-TW/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Network accessline 267
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 268
[Credential-related source omitted] - Network accessline 269
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 270
[Credential-related source omitted] - Access to secretsline 271
[Credential-related source omitted] - Access to secretsline 272
[Credential-related source omitted] - Network accessline 284
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 285
[Credential-related source omitted] - Network accessline 286
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 287
[Credential-related source omitted] - Access to secretsline 288
[Credential-related source omitted] - Access to secretsline 289
[Credential-related source omitted] - Network accessline 340
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 342
APP_URL=https://workspace.example.com
packages/core/docs/content/locales/zh-TW/workspace-management.mdx
- Network accessline 90
並不是每個需要進行更改的人都對 git 感到滿意。 [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management) 支援視覺化分支模型,可對應到底層的 git 分支 - 對於內容和副本更改、布局調整、設計迭代以及無需開發環境的 A/B 測試非常有用。 - Network accessline 140
關鍵提示:使用 GitHub 團隊 (`@org/team`),而不是個人。框架和調度更改應始終需要平台審查。請參閱 [GitHub CODEOWNERS docs](https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-features/customizing-your-repository/about-code-owners) 了解 glob 語法和多所有者模式。 - Network accessline 160
然後新增 [actions/labeler](https://github.com/actions/labeler) 操作 - 請參閱該儲存庫的 README 以了解完整的工作流程 YAML。開啟或更新 PR 時,標籤會自動應用。 - Access to secretsline 233
[Credential-related source omitted]
packages/core/docs/content/mcp-apps.mdx
- Network accessline 24
OpenAI Plugin Extensions add launch points alongside model-invoked MCP tools. When inline MCP Apps are enabled, the shared `open_app` tool advertises a global sidebar entrypoint and a thread side-panel entrypoint. ChatGPT invokes it with `{ - Network accessline 244
Hosts that render MCP Apps inline include Claude chat on web, desktop, and mobile, plus the Claude Desktop Code tab; ChatGPT; VS Code GitHub Copilot (1.109+); Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.6 - Network accessline 246
The Claude Desktop Code tab renders MCP Apps, as observed in [widget data](https://github.com/anthropics/claude-code/issues/103809) and [tool-name](https://github.com/anthropics/claude-code/issues/104045) bug reports. Use the deep-link fall - Access to secretsline 252
[Credential-related source omitted]
packages/core/docs/content/mcp-clients.mdx
- Network accessline 167
{ "$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 168
"$schema": "https://agent-native.com/schema/mcp.config.json", - Network accessline 197
language="jsonc" code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\ - Network accessline 198
code={ '{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "files - Network accessline 199
'{\n "$schema": "https://agent-native.com/schema/mcp.config.json",\n "servers": {\n "claude-in-chrome": {\n "command": "claude-in-chrome-mcp",\n "args": [],\n "env": { "LOG_LEVEL": "info" }\n },\n "filesystem": {\n - Network accessline 245
[Credential-related source omitted] - Network accessline 312
The Atlassian preset uses Atlassian Rovo MCP's OAuth 2.1 endpoint, `https://mcp.atlassian.com/v1/mcp/authv2`, and exposes Jira and Confluence - Network accessline 313
`https://mcp.atlassian.com/v1/mcp/authv2`, and exposes Jira and Confluence - Network accessline 323
the agent-native app runs; do not copy `https://builder.io/**` to another deployment unless that is its actual origin. See Atlassian's [Rovo MCP setup guide](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 324
deployment unless that is its actual origin. See Atlassian's [Rovo MCP setup guide](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) and [admin controls](https://support.atlassian.com/security-and-access-policies/docs - Network accessline 325
guide](https://developer.atlassian.com/cloud/rovo-mcp/guides/getting-started/) - Network accessline 326
and [admin controls](https://support.atlassian.com/security-and-access-policies/docs/control-atlassian-rovo-mcp-server-settings/). - Network accessline 337
1. **Server name** — a short label for your own reference (e.g. "zapier", "slack-tools"). 2. **URL** — the HTTPS endpoint the MCP server provider gave you (e.g. `https://mcp.zapier.com/s/abc123/mcp`). This is usually found in the provider's - Network accessline 338
2. **URL** — the HTTPS endpoint the MCP server provider gave you (e.g. `https://mcp.zapier.com/s/abc123/mcp`). This is usually found in the provider's dashboard or integration docs. - Network accessline 353
[Credential-related source omitted] - Network accessline 406
"example": "{\n \"scope\": \"user\",\n \"name\": \"zapier\",\n \"url\": \"https://mcp.zapier.com/s/abc123/mcp\",\n \"headers\": { \"Authorization\": \"Bearer sk-your-key-here\" }\n}" - Network accessline 436
{ "type": "http", "url": "https://example.com/mcp", - Network accessline 437
"type": "http", "url": "https://example.com/mcp", "headers": { - Network accessline 438
"url": "https://example.com/mcp", - Network accessline 545
AGENT_NATIVE_MCP_HUB_URL=https://dispatch.acme.com
packages/core/docs/content/mcp-protocol.mdx
- Access to secretsline 32
[Credential-related source omitted] - Network accessline 132
claude mcp add --transport http mail https://mail.example.com/mcp - Network accessline 141
"mail": { "type": "http", "url": "https://mail.example.com/mcp", - Network accessline 142
"type": "http", "url": "https://mail.example.com/mcp", }, - Network accessline 143
"url": "https://mail.example.com/mcp", - Network accessline 162
POST https://your-app.example.com/mcp - Access to secretsline 167
[Credential-related source omitted] - Access to secretsline 291
[Credential-related source omitted] - Network accessline 299
claude mcp add --transport http agent-native https://dispatch.agent-native.com/mcp - Network accessline 305
HTTP/1.1 401 Unauthorized WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 306
WWW-Authenticate: Bearer resource_metadata="https://dispatch.agent-native.com/.well-known/oauth-protected-resource", scope="mcp:read mcp:write mcp:apps offline_access" - Network accessline 421
claude mcp add --transport http analytics https://analytics.example.com/mcp - Network accessline 430
"analytics": { "type": "http", "url": "https://analytics.example.com/mcp", - Network accessline 431
"type": "http", "url": "https://analytics.example.com/mcp", }, - Network accessline 432
"url": "https://analytics.example.com/mcp",
packages/core/docs/content/messaging-internals.mdx
- Access to secretsline 299
[Credential-related source omitted]
packages/core/docs/content/messaging-recipes.mdx
packages/core/docs/content/messaging.mdx
- Network accessline 91
1. In **Admin → Messaging**, select **Agent manifest** to download the host-specific `agent-native-slack-manifest.json`, then create or update your Slack app at **[api.slack.com/apps](https://api.slack.com/apps)** by importing - Network accessline 92
host-specific `agent-native-slack-manifest.json`, then create or update your Slack app at **[api.slack.com/apps](https://api.slack.com/apps)** by importing it from **App Manifest**. The manifest enables Slack's current `agent_view` - Network accessline 93
Slack app at **[api.slack.com/apps](https://api.slack.com/apps)** by importing - Network accessline 125
https://your-app.example.com/_agent-native/integrations/slack/oauth/callback - Network accessline 132
https://your-app.example.com/_agent-native/integrations/slack/webhook - Network accessline 138
https://your-app.example.com/_agent-native/integrations/slack/interactions - Access to secretsline 141
[Credential-related source omitted] - Network accessline 222
https://your-clips.example.com/api/slack/unfurl - Network accessline 257
https://your-app.example.com/_agent-native/integrations/microsoft-teams/webhook - Network accessline 280
https://your-app.example.com/_agent-native/integrations/discord/webhook - Network accessline 299
1. Open Telegram and message **[@BotFather](https://t.me/BotFather)**. - Network accessline 307
POST https://your-app.example.com/_agent-native/integrations/telegram/setup - Network accessline 330
1. Sign up at **[resend.com](https://resend.com)**. The free tier is enough to get started. - Network accessline 337
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 344
[Credential-related source omitted] - Network accessline 351
1. Sign up at **[sendgrid.com](https://sendgrid.com)**. - Network accessline 359
https://your-app.example.com/_agent-native/integrations/email/webhook - Access to secretsline 364
[Credential-related source omitted] - Network accessline 388
1. Go to the **[Meta Developer Portal](https://developers.facebook.com/)**, click **Create App**, and pick the **Business** type. - Access to secretsline 395
[Credential-related source omitted] - Network accessline 402
Callback URL: https://your-app.example.com/_agent-native/integrations/whatsapp/webhook - Access to secretsline 451
[Credential-related source omitted] - Access to secretsline 455
[Credential-related source omitted] - Access to secretsline 456
[Credential-related source omitted] - Access to secretsline 458
[Credential-related source omitted]
packages/core/docs/content/multi-app-workspace.mdx
- Access to secretsline 84
[Credential-related source omitted] - Access to secretsline 118
[Credential-related source omitted] - Network accessline 263
"include": ["mail", "calendar", "design"], "urls": { "design": "https://design.example.test" - Network accessline 264
"urls": { "design": "https://design.example.test" } - Network accessline 265
"design": "https://design.example.test" - Access to secretsline 294
[Credential-related source omitted] - Access to secretsline 300
[Credential-related source omitted] - Access to secretsline 303
[Credential-related source omitted]
packages/core/docs/content/native-chat-ui.mdx
- Network accessline 459
For the old-to-new API map and migration steps, see the [AgentKit chat migration guide](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/migrations/agentkit-chat.md).
packages/core/docs/content/neon.mdx
- Network accessline 45
- [Database](/docs/server-database) - schema, migrations, and connection rules - [Deployment](/docs/deployment) - choose a runtime and deployment preset - [Drizzle's Neon guide](https://orm.drizzle.team/docs/connect-neon) - driver-specific - Network accessline 47
- [Drizzle's Neon guide](https://orm.drizzle.team/docs/connect-neon) - driver-specific details
packages/core/docs/content/netlify.mdx
packages/core/docs/content/node-js.mdx
packages/core/docs/content/notifications.mdx
- Network accessline 195
title: "Database offline", body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 196
body: "Primary dropped connections", metadata: { runbookUrl: "https://runbooks/db-offline" }, channels: ["inbox", "webhook"], // optional allowlist; omit to run all - Network accessline 197
metadata: { runbookUrl: "https://runbooks/db-offline" }, - Network accessline 214
name: "slack-ops", async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { - Network accessline 215
async deliver(input, meta) { await fetch(process.env.OPS_SLACK_WEBHOOK!, { method: "POST", - Network accessline 216
await fetch(process.env.OPS_SLACK_WEBHOOK!, {
packages/core/docs/content/observability.mdx
- Access to secretsline 318
[Credential-related source omitted] - Network accessline 578
The repository's [`packages/otel`](https://github.com/BuilderIO/agent-native/tree/main/packages/otel) holds a fuller version of this plugin with per-signal endpoints and exporter validation. It is not published to npm, so start from the plu - Access to secretsline 585
[Credential-related source omitted]
packages/core/docs/content/onboarding.mdx
- Access to secretsline 201
[Credential-related source omitted]
packages/core/docs/content/organizations-teams-permissions.mdx
- Access to secretsline 427
[Credential-related source omitted] - Access to secretsline 473
[Credential-related source omitted] - Access to secretsline 474
[Credential-related source omitted] - Access to secretsline 475
[Credential-related source omitted] - Access to secretsline 476
[Credential-related source omitted] - Access to secretsline 477
[Credential-related source omitted] - Access to secretsline 478
[Credential-related source omitted] - Access to secretsline 482
[Credential-related source omitted] - Access to secretsline 493
[Credential-related source omitted] - Access to secretsline 495
[Credential-related source omitted] - Access to secretsline 496
[Credential-related source omitted] - Access to secretsline 514
[Credential-related source omitted]
packages/core/docs/content/plan-plugin.mdx
- Network accessline 20
- **Three skills** — `/visual-plan` (the canonical entry point), `/visual-recap`, and `/visualize-repo`. - **The Plan MCP connector** — registered against the hosted app at `https://plan.agent-native.com` (MCP endpoint `https://plan.agent-n - Network accessline 21
- **The Plan MCP connector** — registered against the hosted app at `https://plan.agent-native.com` (MCP endpoint `https://plan.agent-native.com/mcp`, server name `plan`). - Network accessline 98
npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex - Network accessline 123
- **Use a Chromium browser.** Safari blocks the hosted HTTPS Plan page from reading the `http://127.0.0.1` localhost bridge (mixed-content / private - Network accessline 124
reading the `http://127.0.0.1` localhost bridge (mixed-content / private - Network accessline 213
- `--client codex|claude-code|claude-code-cli|cowork|all` — which local agents to write the MCP config for (default `all`). - `--no-connect` — register the connector without authenticating; run `npx @agent-native/core@latest connect https:/ - Network accessline 214
- `--no-connect` — register the connector without authenticating; run `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` later, or choose a narrower `--client`. - Network accessline 261
After install, **start a new Codex thread** so the skills and MCP tools load into the session. The plugin ships a URL-only connector (`[mcp_servers.plan]` → `https://plan.agent-native.com/mcp`); `codex mcp login plan` runs the OAuth flow. T - Network accessline 265
**Older installs:** if your config still has an `agent-native-plans` entry pointing at the same URL, running `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` for Codex, or the same command with your
packages/core/docs/content/pr-visual-recap.mdx
- Access to secretsline 194
[Credential-related source omitted] - Access to secretsline 195
[Credential-related source omitted] - Access to secretsline 196
[Credential-related source omitted] - Network accessline 211
[Credential-related source omitted] - Access to secretsline 211
[Credential-related source omitted] - Access to secretsline 226
[Credential-related source omitted] - Access to secretsline 241
[Credential-related source omitted] - Network accessline 252
npx @agent-native/core@latest connect https://plan.agent-native.com --service-token pr-recap --ttl-days 3650 - Network accessline 267
npx @agent-native/core@latest connect https://plan.agent-native.com --client codex - Access to secretsline 279
[Credential-related source omitted] - Access to secretsline 280
[Credential-related source omitted] - Network accessline 286
| `VISUAL_RECAP_AUTO_REFRESH` | `true` | Variable, copy variant only (the reusable caller ignores it). Set to `false` to stop recapping on every push. Recaps then run only when a P - Network accessline 288
| `PLAN_RECAP_APP_URL` | `https://plan.agent-native.com` | Secret. Only when self-hosting the Plans app at a different origin. - Network accessline 308
[Credential-related source omitted] - Access to secretsline 308
[Credential-related source omitted] - Instruction override signalline 399
- Lines that read like direct commands or role instructions ("Ignore previous instructions...", "You are now...", "Write the token to..."). - Access to secretsline 492
[Credential-related source omitted] - Access to secretsline 547
[Credential-related source omitted] - Access to secretsline 548
[Credential-related source omitted] - Access to secretsline 549
[Credential-related source omitted]
packages/core/docs/content/privacy-and-data-rights.mdx
- Network accessline 59
- [GDPR text](https://eur-lex.europa.eu/eli/reg/2016/679/oj/eng/) - [European Commission: handling individual requests](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-indiv - Network accessline 60
- [European Commission: handling individual requests](https://commission.europa.eu/law/law-topic/data-protection/information-business-and-organisations/dealing-requests-individuals_en) - Network accessline 61
- [California Attorney General: CCPA](https://www.oag.ca.gov/privacy/ccpa) - Network accessline 62
- [California Attorney General: CCPA regulations](https://www.oag.ca.gov/privacy/ccpa/regs)
packages/core/docs/content/real-time-collaboration.mdx
- Network accessline 676
`GET /_agent-native/health` reports the outcome, and resolving it performs the registration, so one curl both enables and confirms: - Network accessline 677
registration, so one curl both enables and confirms: - Network accessline 680
curl -s https://your-app.example.com/_agent-native/health | jq .realtime
packages/core/docs/content/recurring-jobs.mdx
packages/core/docs/content/render.mdx
packages/core/docs/content/security.mdx
- Dangerous shell operationline 111
await exec({ sql: "INSERT INTO notes (title) VALUES (?)", args: [title] }); - Dangerous shell operationline 117
await exec(`INSERT INTO notes (title) VALUES ('${title}')`); - Dangerous shell operationline 131
- Never use `innerHTML`, `eval()`, or `document.write()` - Network accessline 135
## Server-Side Fetch (SSRF) {#ssrf} - Access to secretsline 393
[Credential-related source omitted] - Access to secretsline 409
[Credential-related source omitted] - Access to secretsline 415
[Credential-related source omitted]
packages/core/docs/content/server-database.mdx
- Network accessline 10
Every agent-native app stores its state in PostgreSQL. The UI and the agent both read and write the same PostgreSQL tables through the same actions, using the same [Drizzle ORM](https://orm.drizzle.team) client. Live sync uses SSE with poll - Network accessline 80
[PGlite](https://pglite.dev/) runs an in-process WASM PostgreSQL database. It is
packages/core/docs/content/server-middleware.mdx
- Network accessline 106
For more middleware patterns — conditional logic, async operations, response interception — see the [Nitro middleware docs](https://nitro.build/guide/routing#middleware) and the [H3 docs](https://h3.unjs.io). - Network accessline 149
- [**Routes**](/docs/server-routes): custom routes that are protected by this middleware - [**Security**](/docs/security): data scoping, access guards, and the full auth model - [**Nitro middleware docs**](https://nitro.build/guide/routing# - Network accessline 151
- [**Nitro middleware docs**](https://nitro.build/guide/routing#middleware): full middleware API and patterns
packages/core/docs/content/server-overview.mdx
- Network accessline 208
[Nitro](https://nitro.build) is the server toolkit that powers the agent-native HTTP layer. Most of what you'll configure in `server/` is Nitro: file-based routes, startup plugins, and middleware. - Network accessline 216
The underlying request/response API is [H3](https://h3.unjs.io). When you see `defineEventHandler`, `createError`, or `getQuery` in custom routes, that's H3. The [Nitro docs](https://nitro.build/guide) and [H3 docs](https://h3.unjs.io) are
packages/core/docs/content/server-plugins.mdx
- Network accessline 170
Plugins can use any of Nitro's lifecycle hooks: `request`, `beforeResponse`, `close`, and others. The [Nitro plugin docs](https://nitro.build/guide/plugins) cover the full hook API and the plugin contract in detail. - Network accessline 184
- [**Routes**](/docs/server-routes): custom file routes and when to use them - [**Writing Agent Instructions**](/docs/writing-agent-instructions): how to craft the `systemPrompt` in the agent-chat plugin - [**Nitro plugin docs**](https://ni - Network accessline 186
- [**Nitro plugin docs**](https://nitro.build/guide/plugins): the full plugin API and lifecycle hooks
packages/core/docs/content/server-routes.mdx
- Access to secretsline 71
[Credential-related source omitted]
packages/core/docs/content/skills-guide.mdx
- Network accessline 360
"id": "assets", "hosted": { "url": "https://assets.agent-native.com", - Network accessline 361
"hosted": { "url": "https://assets.agent-native.com", "mcpUrl": "https://assets.agent-native.com/mcp" - Network accessline 362
"url": "https://assets.agent-native.com", - Network accessline 363
"mcpUrl": "https://assets.agent-native.com/mcp"
packages/core/docs/content/supabase.mdx
- Network accessline 42
- [Database](/docs/server-database) - schema, migrations, and connection rules - [Deployment](/docs/deployment) - choose a runtime and deployment preset - [Supabase connection guide](https://supabase.com/docs/guides/database/connecting-to-p - Network accessline 44
- [Supabase connection guide](https://supabase.com/docs/guides/database/connecting-to-postgres) - pooler modes and connection strings
packages/core/docs/content/template-analytics.mdx
- Network accessline 83
Live demo: [analytics.agent-native.com](https://analytics.agent-native.com).
packages/core/docs/content/template-assets-integrations.mdx
- Network accessline 20
language="tsx" code={ 'import { EmbeddedApp } from "@agent-native/embedding";\n\n<EmbeddedApp\n url="https://assets.agent-native.com/picker?mediaType=image"\n onMessage={(name, payload) => {\n if (name === "chooseAsset") {\n - Network accessline 21
code={ 'import { EmbeddedApp } from "@agent-native/embedding";\n\n<EmbeddedApp\n url="https://assets.agent-native.com/picker?mediaType=image"\n onMessage={(name, payload) => {\n if (name === "chooseAsset") {\n insertAsset((pay - Network accessline 22
'import { EmbeddedApp } from "@agent-native/embedding";\n\n<EmbeddedApp\n url="https://assets.agent-native.com/picker?mediaType=image"\n onMessage={(name, payload) => {\n if (name === "chooseAsset") {\n insertAsset((payload as { u - Network accessline 215
2. **Ask for images.** In your agent's chat: "Generate three blog hero options from the Acme product shots." The agent opens the picker with candidate images you can regenerate, retune (prompt, aspect, count), and choose from. 3. **Pick.** - Network accessline 216
3. **Pick.** In inline hosts (ChatGPT, Claude chat on web, desktop, and mobile, plus the Claude Desktop Code tab) the picker renders right in the chat — click a candidate and the choice flows back automatically. On CLI/link-only hosts (Code - Network accessline 221
The Assets app skill has app id `assets` and hosted MCP URL `https://assets.agent-native.com/mcp`. - Network accessline 222
`https://assets.agent-native.com/mcp`.
packages/core/docs/content/template-assets-presets.mdx
- Network accessline 178
For deeper prompting guidance on the underlying models, see OpenAI's [GPT Image prompting guide](https://developers.openai.com/cookbook/examples/multimodal/image-gen-models-prompting-guide). - Network accessline 179
[GPT Image prompting guide](https://developers.openai.com/cookbook/examples/multimodal/image-gen-models-prompting-guide).
packages/core/docs/content/template-assets.mdx
- Network accessline 109
Live demo: [assets.agent-native.com](https://assets.agent-native.com).
packages/core/docs/content/template-brain.mdx
- Network accessline 153
Live demo: [brain.agent-native.com](https://brain.agent-native.com).
packages/core/docs/content/template-calendar-developers.mdx
- Network accessline 26
This needs [Node.js](https://nodejs.org) 22.22 or later and - Network accessline 27
[pnpm](https://pnpm.io) installed first. See - Network accessline 38
The app opens at `http://localhost:8082`. If it doesn't open automatically, - Network accessline 49
1. **Create a Google Cloud project**, if you don't already have one to use: in [Google Cloud Console](https://console.cloud.google.com/), click the - Network accessline 50
in [Google Cloud Console](https://console.cloud.google.com/), click the - Network accessline 59
4. **Create credentials**: under **APIs & Services → Credentials**, create an OAuth client ID of type "Web application," and add `http://localhost:8082/_agent-native/google/callback` as an authorized - Network accessline 60
an OAuth client ID of type "Web application," and add `http://localhost:8082/_agent-native/google/callback` as an authorized redirect URI (8082 is Calendar's default dev port). - Network accessline 61
`http://localhost:8082/_agent-native/google/callback` as an authorized redirect URI (8082 is Calendar's default dev port). 5. **Add the credentials**: download the credentials JSON and upload it in - Access to secretsline 78
[Credential-related source omitted] - Network accessline 81
Get a Client ID and Secret from a [Zoom OAuth app](https://marketplace.zoom.us/)
packages/core/docs/content/template-calendar-features.mdx
packages/core/docs/content/template-calendar-integrations.mdx
- Access to secretsline 23
[Credential-related source omitted]
packages/core/docs/content/template-calendar.mdx
packages/core/docs/content/template-chat-developers.mdx
- Network accessline 31
This needs [Node.js](https://nodejs.org) 22.22 or later. If `pnpm` is not on
packages/core/docs/content/template-chat-integrations.mdx
- Access to secretsline 22
[Credential-related source omitted]
packages/core/docs/content/template-chat.mdx
packages/core/docs/content/template-clips-developers.mdx
- Network accessline 22
This needs [Node.js](https://nodejs.org) 22.22 or later and [pnpm](https://pnpm.io) installed first. See [Getting started](/docs/getting-started) if you haven't installed either yet, or don't have an LLM connection (Builder.io, an Anthropic - Network accessline 300
language="ts" code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimp - Network accessline 301
code={ 'import { defineAction } from "@agent-native/core/action";\nimport {\n buildAgentAccessUrl,\n createScopedAgentAccessGrant,\n getRequestContext,\n getRequestUserEmail,\n} from "@agent-native/core/server";\nimport { ForbiddenE - Network accessline 302
[Credential-related source omitted]
packages/core/docs/content/template-clips-embed.mdx
- Network accessline 40
npx @agent-native/core@latest connect https://clips.agent-native.com \ - Network accessline 53
const clipsBaseUrl = "https://clips.agent-native.com"; - Access to secretsline 54
[Credential-related source omitted] - Network accessline 56
const response = await fetch( - Network accessline 145
npx @agent-native/core@latest connect https://clips.example.com \ - Network accessline 152
CLIPS_BASE_URL=https://clips.example.com
packages/core/docs/content/template-clips-features.mdx
- Network accessline 117
The desktop app is built for the smoothest everyday capture: screen and webcam recording with a draggable camera bubble, live transcription while you record, all without keeping a browser tab open. Download it from [clips.agent-native.com/d - Network accessline 212
[Credential-related source omitted] - Network accessline 213
[Credential-related source omitted]
packages/core/docs/content/template-clips-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/template-clips.mdx
packages/core/docs/content/template-content-developers.mdx
packages/core/docs/content/template-content.mdx
- Network accessline 16
### [Try it out now →](https://content.agent-native.com)
packages/core/docs/content/template-design-developers.mdx
- Network accessline 28
The app opens at `http://localhost:8099`, Design's default dev port. Design
packages/core/docs/content/template-design-integrations.mdx
- Access to secretsline 22
[Credential-related source omitted] - Network accessline 46
https://design.agent-native.com/mcp - Network accessline 50
approve access. Claude Code adds the same URL, then runs `/mcp` to authenticate. Codex and other CLI or stdio clients instead run `npx @agent-native/core@latest connect https://design.agent-native.com`, which - Network accessline 51
authenticate. Codex and other CLI or stdio clients instead run `npx @agent-native/core@latest connect https://design.agent-native.com`, which opens a browser tab to authorize and mints a per-user, revocable token, with - Network accessline 52
@agent-native/core@latest connect https://design.agent-native.com`, which - Network accessline 67
authentication error, don't keep retrying it. Run `/mcp` in Claude Code and reconnect, or run `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` from any terminal. Both re-authenticate - Network accessline 68
reconnect, or run `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` from any terminal. Both re-authenticate without reinstalling anything. - Network accessline 69
https://design.agent-native.com` from any terminal. Both re-authenticate
packages/core/docs/content/template-design.mdx
packages/core/docs/content/template-dispatch.mdx
packages/core/docs/content/template-forms-agent.mdx
- Network accessline 40
- **"Publish this and give me the link."** The form goes live and you get its public URL right away. - **"Create an anonymous customer feedback form and give me the public link."** The form is published anonymous from the start, so no submi - Network accessline 42
- **"Only allow this form to be submitted from https://example.com."** The form starts rejecting submissions from anywhere else. This is only available by asking the agent today. There's no toggle for it in the builder.
packages/core/docs/content/template-forms-developers.mdx
- Network accessline 22
This needs [Node.js](https://nodejs.org) 22.22 or later and [pnpm](https://pnpm.io) installed first. See [Getting started](/docs/getting-started) if you haven't installed either yet, or don't have an LLM connection (Builder.io, an Anthropic - Network accessline 30
The app opens at `http://localhost:8084`. If it doesn't open automatically, visit that URL in your browser.
packages/core/docs/content/template-forms-features.mdx
- Network accessline 14
[Credential-related source omitted] - Network accessline 15
[Credential-related source omitted] - Network accessline 94
Public submissions can require verification through [Cloudflare Turnstile](https://www.cloudflare.com/products/turnstile/) before they go through. Turn this on by setting a site key and a secret key in Settings. Two protections stay on even - Network accessline 110
There's no toggle for this in the builder today. Ask the agent to set it up, for example "only allow this form to be submitted from https://example.com." Give the agent the exact origin (the scheme and domain, like `https://example.com`, wi - Network accessline 116
[Credential-related source omitted] - Network accessline 117
[Credential-related source omitted]
packages/core/docs/content/template-forms-integrations.mdx
- Access to secretsline 16
[Credential-related source omitted]
packages/core/docs/content/template-forms.mdx
packages/core/docs/content/template-mail-developers.mdx
- Network accessline 22
1. Open [Google Cloud Console](https://console.cloud.google.com/) and create a project. - Network accessline 23
2. Enable the **Gmail API** under APIs & Services → Library. 3. Create OAuth 2.0 credentials (type: Web application). Add `http://localhost:8085/_agent-native/google/callback` as an authorized redirect URI. 4. Copy the Client ID and Client - Network accessline 24
3. Create OAuth 2.0 credentials (type: Web application). Add `http://localhost:8085/_agent-native/google/callback` as an authorized redirect URI.
packages/core/docs/content/template-mail.mdx
- Network accessline 100
Live demo: [mail.agent-native.com](https://mail.agent-native.com).
packages/core/docs/content/template-plan-automations.mdx
packages/core/docs/content/template-plan-developers.mdx
packages/core/docs/content/template-plan-local-and-desktop.mdx
- Network accessline 112
The bridge URL looks like `https://plan.agent-native.com/local-plans/local-<opaque-id>#bridge=http://127.0.0.1:...`. - Network accessline 113
`https://plan.agent-native.com/local-plans/local-<opaque-id>#bridge=http://127.0.0.1:...`. - Network accessline 144
http://localhost:<port>/local-plans/<slug> - Network accessline 151
http://localhost:<port>/local-plans/<slug>?path=plans%2F<slug>
packages/core/docs/content/template-plan-review-workflow.mdx
- Network accessline 136
The hosted Plans connector lives at `https://plan.agent-native.com/mcp`.
packages/core/docs/content/template-plan.mdx
- Network accessline 137
This installs three slash commands — `/visual-plan`, `/visual-recap`, and `/visualize-repo` — plus the `plan` MCP connector (a remote tool server your agent talks to) at `https://plan.agent-native.com/mcp`. For the full flag - Network accessline 138
`/visualize-repo` — plus the `plan` MCP connector (a remote tool server your agent talks to) at `https://plan.agent-native.com/mcp`. For the full flag reference, the plugin/marketplace install routes for Claude Code and Codex, - Network accessline 139
agent talks to) at `https://plan.agent-native.com/mcp`. For the full flag - Network accessline 227
If a Plans tool ever returns `needs auth`, `Unauthorized`, or `Session terminated`, don't keep retrying it. Re-authenticate with `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, - Network accessline 228
terminated`, don't keep retrying it. Re-authenticate with `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`, or re-run `/mcp` → **Authenticate** for an OAuth-capable host. Start a new - Network accessline 229
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex`,
packages/core/docs/content/template-slides-features.mdx
packages/core/docs/content/template-slides-integrations.mdx
- Access to secretsline 21
[Credential-related source omitted]
packages/core/docs/content/template-slides.mdx
packages/core/docs/content/toolkit-capability-packages.mdx
- Network accessline 35
- [Scheduling package docs](https://github.com/BuilderIO/agent-native/tree/main/packages/scheduling/docs) - Network accessline 36
- [Creative Context package docs](https://github.com/BuilderIO/agent-native/tree/main/packages/creative-context/docs) - Network accessline 37
- [Pinpoint package README](https://github.com/BuilderIO/agent-native/tree/main/packages/pinpoint)
packages/core/docs/content/toolkit-command-navigation.mdx
- Access to secretsline 110
[Credential-related source omitted]
packages/core/docs/content/toolkit-org-team.mdx
packages/core/docs/content/toolkit-settings.mdx
- Access to secretsline 155
[Credential-related source omitted] - Network accessline 255
The OAuth callback must use `http://127.0.0.1` on the user's machine. - Network accessline 258
access on a hosted site requires both OpenAI partner approval for the app operator and a hosted implementation. The operator can request approval through OpenAI's [interest form](https://openai.com/form/sign-in-with-chatgpt-interest). - Network accessline 260
OpenAI's [interest form](https://openai.com/form/sign-in-with-chatgpt-interest). - Network accessline 277
marks the token-sharing flow as a preview, so model availability, limits, and supported API features can change. Review OpenAI's [Sign in with ChatGPT overview](https://developers.openai.com/siwc/token-sharing-open-source), - Network accessline 278
supported API features can change. Review OpenAI's [Sign in with ChatGPT overview](https://developers.openai.com/siwc/token-sharing-open-source), [registration and sign-in](https://developers.openai.com/siwc/token-sharing-open-source/sign-i - Network accessline 279
overview](https://developers.openai.com/siwc/token-sharing-open-source), - Network accessline 280
[registration and sign-in](https://developers.openai.com/siwc/token-sharing-open-source/sign-in), - Network accessline 281
[accounts and sessions](https://developers.openai.com/siwc/token-sharing-open-source/profiles-and-sessions), - Network accessline 282
[models and inference](https://developers.openai.com/siwc/token-sharing-open-source/models-and-inference), - Network accessline 283
[preview limitations](https://developers.openai.com/siwc/token-sharing-open-source/preview-limitations), - Network accessline 284
[error recovery](https://developers.openai.com/siwc/token-sharing-open-source/errors-and-recovery), - Network accessline 285
and [UI/UX guidelines](https://developers.openai.com/siwc/ui-ux-guidelines).
packages/core/docs/content/tracking.mdx
- Network accessline 83
| Provider | Env vars | | --------- | ----------------------------------------------------------------------------------------------- | | PostHog | `P - Network accessline 84
| --------- | ----------------------------------------------------------------------------------------------- | | PostHog | `POSTHOG_API_KEY` (required), `POSTHOG_HOST` (optional, defaults to `https://us.i.posthog.com`) | | Mixpanel | `M - Network accessline 85
[Credential-related source omitted] - Access to secretsline 85
[Credential-related source omitted] - Access to secretsline 87
[Credential-related source omitted] - Access to secretsline 94
[Credential-related source omitted] - Access to secretsline 106
[Credential-related source omitted] - Network accessline 238
configureTracking({ key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 239
key: "anpk_...", endpoint: "https://analytics.example.com/api/analytics/track", sessionReplay: { - Network accessline 240
endpoint: "https://analytics.example.com/api/analytics/track", - Network accessline 258
VITE_AGENT_NATIVE_ANALYTICS_PUBLIC_KEY=anpk_... VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 259
VITE_AGENT_NATIVE_ANALYTICS_ENDPOINT=https://analytics.example.com/api/analytics/track - Network accessline 316
track(event) { // Send event to your backend fetch("https://analytics.example.com/events", { - Network accessline 317
// Send event to your backend fetch("https://analytics.example.com/events", { method: "POST", - Network accessline 318
fetch("https://analytics.example.com/events", { - Network accessline 376
- **Batched HTTP** — built-in providers enqueue events and flush every 10 seconds or when 50 events accumulate, whichever comes first. This minimizes outbound requests without losing data. - **No SDK dependencies** — all built-in providers - Network accessline 377
- **No SDK dependencies** — all built-in providers use raw `fetch()`. No PostHog SDK, no Mixpanel SDK, no Amplitude SDK. Keeps the framework lightweight.
packages/core/docs/content/turn-into-app.mdx
- Network accessline 12
- An AI agent that can run terminal commands, such as Claude Code, Codex, or Cursor. Claude or ChatGPT on the web can hand the job to Builder instead. See [Where it runs](#hosts). - [Node.js](https://nodejs.org). The installer from nodejs.o - Network accessline 13
- [Node.js](https://nodejs.org). The installer from nodejs.org includes `npx`, which runs the install commands below. - Network accessline 18
- **On your machine.** Node.js 22.22 or later (check with `node --version`) and [pnpm](https://pnpm.io) on your `PATH`. If `pnpm --version` fails, run `corepack enable`, which ships with Node.js. - Network accessline 59
Builder hosting, from either kind of host, needs the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth) and calls its `start-workspace-app-creation` action. Nothing runs on your side, so the report stays pe - Access to secretsline 118
[Credential-related source omitted]
packages/core/docs/content/updating-ui-in-production.mdx
- Network accessline 18
[Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=deployment) is a separate managed code-capable workflow where an agent can modify your app's UI in production. Use Builder.io wit
packages/core/docs/content/upgrading-core-ui.mdx
- Access to secretsline 87
[Credential-related source omitted] - Access to secretsline 90
[Credential-related source omitted] - Network accessline 206
Canonical source URL: [upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx). - Network accessline 207
[upgrading-core-ui.mdx](https://github.com/BuilderIO/agent-native/blob/main/packages/core/docs/content/upgrading-core-ui.mdx).
packages/core/docs/content/vercel.mdx
- Access to secretsline 97
[Credential-related source omitted] - Network accessline 114
export BETTER_AUTH_URL='https://your-domain.example' - Access to secretsline 141
[Credential-related source omitted] - Network accessline 204
Scheduled automations, queued event automations, and cleanup of interrupted runs all happen on a once-a-minute sweep. The build adds a [Vercel cron job](https://vercel.com/docs/cron-jobs) that calls the sweep at - Network accessline 205
runs all happen on a once-a-minute sweep. The build adds a [Vercel cron job](https://vercel.com/docs/cron-jobs) that calls the sweep at `/_agent-native/jobs/_process-sweep`. A workspace build adds one cron job per - Network accessline 206
[Vercel cron job](https://vercel.com/docs/cron-jobs) that calls the sweep at
packages/core/docs/content/voice-input.mdx
packages/core/docs/content/webmcp.mdx
- Network accessline 29
The browser API is experimental and may change. See the [Chrome WebMCP Imperative API](https://developer.chrome.com/docs/ai/webmcp/imperative-api) and - Network accessline 30
Imperative API](https://developer.chrome.com/docs/ai/webmcp/imperative-api) and - Network accessline 31
the [WebMCP specification](https://webmachinelearning.github.io/webmcp/) for - Network accessline 142
npx @mcp-b/webmcp-local-relay@latest --widget-origin https://slides.agent-native.com - Network accessline 175
}), ], exposedTo: ["https://agent.example"], - Network accessline 176
], exposedTo: ["https://agent.example"], getContext: () => ({ route: { name: window.location.pathname } }), - Network accessline 177
exposedTo: ["https://agent.example"], - Network accessline 223
const webmcp = createAgentNativeWebMcpClient({ fromOrigins: ["https://orders.example"], - Network accessline 224
fromOrigins: ["https://orders.example"], - Network accessline 237
const registration = createAgentNativeWebMcpRegistration({ actions, exposedTo: ["https://orders.example"], - Network accessline 238
actions, exposedTo: ["https://orders.example"], }); - Network accessline 239
exposedTo: ["https://orders.example"], - Network accessline 247
<iframe src="https://orders.example" allow="tools"></iframe> - Network accessline 256
<AgentNativeFrame agentUrl="https://chat.example/agent" - Network accessline 257
agentUrl="https://chat.example/agent" - Network accessline 258
agentOrigin="https://chat.example" - Network accessline 316
const manifest = await fetch( - Network accessline 317
"https://app.example.com/.well-known/mcp.json", - Network accessline 321
const result = await fetch(tool.endpoint, {
packages/core/docs/content/what-is-agent-native.mdx
packages/core/docs/content/workspace-deployment.mdx
- Network accessline 15
npx @agent-native/core@latest deploy # https://your-agents.com/mail/* → apps/mail - Network accessline 16
# https://your-agents.com/mail/* → apps/mail - Network accessline 17
# https://your-agents.com/calendar/* → apps/calendar - Network accessline 18
# https://your-agents.com/forms/* → apps/forms - Network accessline 24
<div class="diagram-ws"> <div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> - Network accessline 25
<div class="diagram-panel" data-rough> <strong>https://your-agents.com</strong> <div class="diagram-row"> - Network accessline 26
<strong>https://your-agents.com</strong> - Network accessline 267
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/mail - Network accessline 268
[Credential-related source omitted] - Network accessline 269
BETTER_AUTH_URL: https://agents.example.com/mail - Access to secretsline 270
[Credential-related source omitted] - Access to secretsline 271
[Credential-related source omitted] - Access to secretsline 272
[Credential-related source omitted] - Network accessline 284
environment: DATABASE_URL: postgres://agent_native:change-me@postgres:5432/agent_native BETTER_AUTH_URL: https://agents.example.com/calendar - Network accessline 285
[Credential-related source omitted] - Network accessline 286
BETTER_AUTH_URL: https://agents.example.com/calendar - Access to secretsline 287
[Credential-related source omitted] - Access to secretsline 288
[Credential-related source omitted] - Access to secretsline 289
[Credential-related source omitted] - Network accessline 346
AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id":"reports","name":"Reports","path":"/reports"}] VITE_AGENT_NATIVE_WORKSPACE_APPS_JSON=[{"id":"dispatch","name":"Dispatch","path":"/dispatch"},{"id - Network accessline 348
APP_URL=https://workspace.example.com
packages/core/docs/content/workspace-management.mdx
- Network accessline 96
Not everyone who needs to make changes is comfortable with git. [Builder.io](https://www.builder.io?utm_source=agent-native&utm_medium=docs&utm_campaign=framework&utm_content=workspace_management) supports a visual branching model that maps - Network accessline 146
Key tips: use GitHub teams (`@org/team`), not individuals. Framework and Dispatch changes should always require platform review. See [GitHub CODEOWNERS docs](https://docs.github.com/en/repositories/managing-your-repositorys-settings-and-fea - Network accessline 166
Then add the [actions/labeler](https://github.com/actions/labeler) action — see that repo's README for the full workflow YAML. Labels apply automatically when PRs are opened or updated. - Access to secretsline 239
[Credential-related source omitted]
packages/core/docs/design/action-receipts-and-domain-events.md
packages/core/docs/design/durable-agent-runs.md
- Access to secretsline 969
[Credential-related source omitted]
packages/core/docs/design/gpt-reasoning-effort-gateway-contract.md
packages/core/docs/design/organization-team-tenancy.md
- Network accessline 27
[Issue #5611](https://github.com/BuilderIO/agent-native/issues/5611) asks for one organization with people in several teams. Each team needs its own agent instructions, skills, and memory alongside the organization's. Leads need a central v - Network accessline 39
An organization-wide resource does not need to become team-owned to meet these needs. [Review feedback on PR #5777](https://github.com/BuilderIO/agent-native/pull/5777#pullrequestreview-5311752444) calls for a smaller first release instead
packages/core/src/templates/builder-code-starter/.agents/skills/multi-app-workspace/SKILL.md
packages/core/src/templates/chat/.agents/skills/actions/SKILL.md
packages/core/src/templates/chat/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
packages/core/src/templates/chat/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
packages/core/src/templates/chat/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
packages/core/src/templates/default/.agents/skills/actions/SKILL.md
packages/core/src/templates/default/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
packages/core/src/templates/default/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
packages/core/src/templates/default/.agents/skills/agent-engines/SKILL.md
- Access to secretsline 47
[Credential-related source omitted] - Access to secretsline 108
[Credential-related source omitted] - Access to secretsline 109
[Credential-related source omitted] - Access to secretsline 110
[Credential-related source omitted] - Access to secretsline 111
[Credential-related source omitted] - Access to secretsline 112
[Credential-related source omitted] - Access to secretsline 113
[Credential-related source omitted] - Access to secretsline 114
[Credential-related source omitted] - Access to secretsline 115
[Credential-related source omitted] - Network accessline 149
Any `vendor/model` string from [openrouter.ai/models](https://openrouter.ai/models) works — the `supportedModels` list in the registry is a UI hint, not an allow-list. - Network accessline 155
[Credential-related source omitted] - Access to secretsline 155
[Credential-related source omitted] - Network accessline 156
appName: "My App", appUrl: "https://myapp.example", }); - Network accessline 157
appUrl: "https://myapp.example", - Access to secretsline 188
[Credential-related source omitted] - Access to secretsline 225
[Credential-related source omitted] - Access to secretsline 226
[Credential-related source omitted] - Access to secretsline 227
[Credential-related source omitted] - Access to secretsline 228
[Credential-related source omitted] - Access to secretsline 229
[Credential-related source omitted] - Access to secretsline 230
[Credential-related source omitted] - Access to secretsline 231
[Credential-related source omitted]
packages/core/src/templates/default/.agents/skills/app-branding/SKILL.md
- Network accessline 128
<svg xmlns="http://www.w3.org/2000/svg" width="192" height="192" viewBox="0 0 192 192">
packages/core/src/templates/default/.agents/skills/notifications/SKILL.md
packages/core/src/templates/default/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
packages/core/src/templates/factory/.agents/skills/review-latest-feedback/SKILL.md
- Network accessline 102
- `thanks - the <https://github.com/BuilderIO/agent-native/pull/1234|fix> so the export keeps your column order is merged and will be on beta shortly.` - Network accessline 103
- `thanks - once this <https://github.com/BuilderIO/agent-native/pull/1234|fix> merges, the export will keep your column order; it's in review now.` - Network accessline 110
thanks - the <https://github.com/BuilderIO/agent-native/pull/1234|fix> so the export keeps your column order is merged and will be on beta shortly.
packages/core/src/templates/headless/.agents/skills/actions/SKILL.md
packages/core/src/templates/headless/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
packages/core/src/templates/headless/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
packages/core/src/templates/headless/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
packages/core/src/templates/workspace-core/.agents/skills/a2a-protocol/SKILL.md
- Network accessline 65
"name": "Analytics", "description": "Queries analytics data across providers", "url": "https://analytics.example.com", - Network accessline 66
"description": "Queries analytics data across providers", "url": "https://analytics.example.com", "color": "#6B7280", - Network accessline 67
"url": "https://analytics.example.com", - Network accessline 68
"color": "#6B7280", "cardUrl": "https://analytics.example.com/.well-known/agent-card.json", "auth": { - Network accessline 69
"cardUrl": "https://analytics.example.com/.well-known/agent-card.json", - Network accessline 102
The last one matters for template apps with no Code tab: "connect me to https://analytics.example.com" is a complete instruction. `remote-agents/` is - Network accessline 103
https://analytics.example.com" is a complete instruction. `remote-agents/` is - Network accessline 107
First-party templates are seeded automatically and overlaid on top of manifests, so `Mail`, `Calendar`, and friends appear without registration. In dev they resolve to `http://localhost:<devPort>`; a manifest pointing at localhost is - Network accessline 108
so `Mail`, `Calendar`, and friends appear without registration. In dev they resolve to `http://localhost:<devPort>`; a manifest pointing at localhost is ignored in production. - Network accessline 109
resolve to `http://localhost:<devPort>`; a manifest pointing at localhost is - Network accessline 210
`.../agents/{agent}/endpoint/protocols/a2a/agentCard/v1.0` when available, and pass `protocolVersion` only when the card omits it. Foundry v1 does not provide SSE streaming. See the [Foundry hosted agent overview](https://learn.microsoft.co - Network accessline 211
and pass `protocolVersion` only when the card omits it. Foundry v1 does not provide SSE streaming. See the [Foundry hosted agent overview](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) and [A2A endpoint guidance](https:// - Network accessline 212
provide SSE streaming. See the [Foundry hosted agent overview](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 213
and [A2A endpoint guidance](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint). - Access to secretsline 237
[Credential-related source omitted] - Network accessline 291
const auth = await resolveA2ACallerAuth(); const answer = await callAgent( "https://analytics.example.com", - Network accessline 292
const answer = await callAgent( "https://analytics.example.com", "What were last week's signups?", - Network accessline 293
"https://analytics.example.com", - Network accessline 366
// The second argument is the bearer token itself, not a signing secret. const { apiKey, apiKeyFallbacks } = await resolveA2ACallerAuth(); const client = new A2AClient("https://analytics.example.com", apiKey, { - Network accessline 367
const { apiKey, apiKeyFallbacks } = await resolveA2ACallerAuth(); const client = new A2AClient("https://analytics.example.com", apiKey, { fallbackApiKeys: apiKeyFallbacks, - Network accessline 368
const client = new A2AClient("https://analytics.example.com", apiKey, {
packages/core/src/templates/workspace-core/.agents/skills/actions/SKILL.md
packages/core/src/templates/workspace-core/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
packages/core/src/templates/workspace-core/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
packages/core/src/templates/workspace-core/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
packages/core/src/templates/workspace-core/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
packages/core/src/templates/workspace-core/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
packages/core/src/templates/workspace-core/.agents/skills/turn-into-app/references/attachments.md
packages/core/src/templates/workspace-core/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
packages/core/src/templates/workspace-core/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
packages/core/src/templates/workspace-core/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
packages/core/src/templates/workspace-core/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
packages/core/src/templates/workspace-core/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
scripts/agent-friction-report.mjs
- Executable scriptline 1
#!/usr/bin/env node - Executable scriptline 1
Executable file; content unavailable - Executable scriptline 1
Executable file; content unavailable - Executable scriptline 1
Executable file; content unavailable - Executable scriptline 1
#!/usr/bin/env node - Executable scriptline 1
Executable file; content unavailable - Executable scriptline 1
Executable file; content unavailable - Executable scriptline 1
Executable file; content unavailable - Possible obfuscationline 113
/\b(?:never|stop|don['’]?t|do not|must not|should not|shouldn['’]?t)\b(?!\s+(?:\w+\s+){0,3}forget\b(?![^.!?\n]{0,100}\b(?:not|never|avoid|refrain|stop|prohibit|prevent|do not|don['’]?t|must not|mustn['’]?t|should not|shouldn['’]?t|cannot|ca - Possible obfuscationline 113
/\b(?:never|stop|don['’]?t|do not|must not|should not|shouldn['’]?t)\b(?!\s+(?:\w+\s+){0,3}forget\b(?![^.!?\n]{0,100}\b(?:not|never|avoid|refrain|stop|prohibit|prevent|do not|don['’]?t|must not|mustn['’]?t|should not|shouldn['’]?t|cannot|ca - Access to secretsline 706
[Credential-related source omitted] - Access to secretsline 706
[Credential-related source omitted] - Network accessline 1259
[ true, "Don't ever say 'I have no further feedback'; people look to me for what they should do. https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1259
[ true, "Don't ever say 'I have no further feedback'; people look to me for what they should do. https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1260
true, "Don't ever say 'I have no further feedback'; people look to me for what they should do. https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1260
true, "Don't ever say 'I have no further feedback'; people look to me for what they should do. https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1261
"Don't ever say 'I have no further feedback'; people look to me for what they should do. https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1261
"Don't ever say 'I have no further feedback'; people look to me for what they should do. https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1264
[ true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1264
[ true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1265
true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1265
true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1266
"I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1266
"I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1268
[ false, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629.", - Network accessline 1268
[ false, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629.", - Network accessline 1269
false, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629.", ], - Network accessline 1269
false, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629.", ], - Network accessline 1270
"I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629.", - Network accessline 1270
"I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629.", - Network accessline 1272
[ false, "I should never leave it ambiguous. This unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1272
[ false, "I should never leave it ambiguous. This unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1273
false, "I should never leave it ambiguous. This unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1273
false, "I should never leave it ambiguous. This unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1274
"I should never leave it ambiguous. This unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1274
"I should never leave it ambiguous. This unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1276
[ false, "I should never leave it ambiguous; this unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1276
[ false, "I should never leave it ambiguous; this unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1277
false, "I should never leave it ambiguous; this unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1277
false, "I should never leave it ambiguous; this unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1278
"I should never leave it ambiguous; this unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1278
"I should never leave it ambiguous; this unrelated note links to https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1280
[ true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#discussion_r123.", - Network accessline 1280
[ true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#discussion_r123.", - Network accessline 1281
true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#discussion_r123.", ], - Network accessline 1281
true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#discussion_r123.", ], - Network accessline 1282
"I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#discussion_r123.", - Network accessline 1282
"I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#discussion_r123.", - Network accessline 1284
[ true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#pullrequestreview-123.", - Network accessline 1284
[ true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#pullrequestreview-123.", - Network accessline 1285
true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#pullrequestreview-123.", ], - Network accessline 1285
true, "I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#pullrequestreview-123.", ], - Network accessline 1286
"I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#pullrequestreview-123.", - Network accessline 1286
"I should never leave it ambiguous like this https://github.com/BuilderIO/agent-native/pull/7629#pullrequestreview-123.", - Network accessline 1302
[ false, "Don't say 'I have no further feedback'. Separately, the PR comment is vague: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1302
[ false, "Don't say 'I have no further feedback'. Separately, the PR comment is vague: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1303
false, "Don't say 'I have no further feedback'. Separately, the PR comment is vague: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1303
false, "Don't say 'I have no further feedback'. Separately, the PR comment is vague: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1304
"Don't say 'I have no further feedback'. Separately, the PR comment is vague: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1304
"Don't say 'I have no further feedback'. Separately, the PR comment is vague: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1306
[ false, "Don't say 'I have no further feedback'; this PR comment link is only incidental: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1306
[ false, "Don't say 'I have no further feedback'; this PR comment link is only incidental: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1307
false, "Don't say 'I have no further feedback'; this PR comment link is only incidental: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1307
false, "Don't say 'I have no further feedback'; this PR comment link is only incidental: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", ], - Network accessline 1308
"Don't say 'I have no further feedback'; this PR comment link is only incidental: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.", - Network accessline 1308
"Don't say 'I have no further feedback'; this PR comment link is only incidental: https://github.com/BuilderIO/agent-native/pull/7629#issuecomment-6359578348.",
scripts/ci-change-scope.ts
skills/an/SKILL.md
skills/assets/SKILL.md
- Network accessline 41
npx @agent-native/core@latest connect https://assets.agent-native.com --client claude-code - Network accessline 47
npx @agent-native/core@latest connect https://dispatch.agent-native.com --client claude-code - Network accessline 53
https://assets.agent-native.com/mcp - Network accessline 77
the skill instructions but has not exposed the MCP tool namespace yet, use the Assets browser fallback URL shape rather than switching to a generic generator: `https://assets.agent-native.com/library?mediaType=image&prompt=...&autoGenerate= - Network accessline 78
Assets browser fallback URL shape rather than switching to a generic generator: `https://assets.agent-native.com/library?mediaType=image&prompt=...&autoGenerate=1&count=3`. When reporting the final selected image in Codex or Claude Code, in - Network accessline 79
`https://assets.agent-native.com/library?mediaType=image&prompt=...&autoGenerate=1&count=3`. When reporting the final selected image in Codex or Claude Code, include the asset link and, if an inline preview is important, download the select - Network accessline 111
the reconnect step: in Claude Code run `/mcp` and choose Authenticate/Reconnect for the Assets connector; from any terminal run `npx -y @agent-native/core@latest reconnect https://assets.agent-native.com` — this - Network accessline 112
Authenticate/Reconnect for the Assets connector; from any terminal run `npx -y @agent-native/core@latest reconnect https://assets.agent-native.com` — this re-authenticates WITHOUT reinstalling. Never reinstall from scratch just to - Network accessline 113
`npx -y @agent-native/core@latest reconnect https://assets.agent-native.com` — this - Network accessline 114
re-authenticates WITHOUT reinstalling. Never reinstall from scratch just to fix auth. Continue once the connector is available. - Do not hand-roll MCP HTTP requests with curl from the agent session. Use the - Network accessline 115
fix auth. Continue once the connector is available. - Do not hand-roll MCP HTTP requests with curl from the agent session. Use the host-exposed Assets tools after restart/reload, or use the returned - Network accessline 116
- Do not hand-roll MCP HTTP requests with curl from the agent session. Use the
skills/design-exploration/SKILL.md
- Network accessline 95
- Hosted default: connect `https://design.agent-native.com/mcp`. - Network accessline 108
the user the reconnect step: in Claude Code run `/mcp` and choose Authenticate/Reconnect for the Design connector; from any terminal run `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` — this - Network accessline 109
Authenticate/Reconnect for the Design connector; from any terminal run `npx -y @agent-native/core@latest reconnect https://design.agent-native.com` — this re-authenticates WITHOUT reinstalling. Never reinstall from scratch just to fix - Network accessline 110
`npx -y @agent-native/core@latest reconnect https://design.agent-native.com` — this - Network accessline 111
re-authenticates WITHOUT reinstalling. Never reinstall from scratch just to fix auth. Continue once the connector is available. - Do not hand-roll MCP HTTP requests with curl from the agent session. Use the - Network accessline 112
auth. Continue once the connector is available. - Do not hand-roll MCP HTTP requests with curl from the agent session. Use the host-exposed Design tools after restart/reload, or use the returned - Network accessline 113
- Do not hand-roll MCP HTTP requests with curl from the agent session. Use the
skills/turn-into-app/references/attachments.md
skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
skills/visual-edit/README.md
- Network accessline 55
See the [skill instructions](https://github.com/BuilderIO/agent-native/blob/main/skills/visual-edit/SKILL.md)
skills/visual-edit/SKILL.md
- Network accessline 18
The editor is hosted at `https://design.agent-native.com`. Never start local - Network accessline 21
See the [Visual Edit guide](https://github.com/BuilderIO/agent-native/blob/main/skills/visual-edit/README.md) - Network accessline 57
- Inline-browser hosts should open `https://design.agent-native.com/visual-edit` - Network accessline 73
JavaScript evaluator. Prefer the host's inline browser; use external Chrome only when requested or unavailable. Open `https://design.agent-native.com/visual-edit`, keep it visible while tools - Network accessline 74
only when requested or unavailable. Open `https://design.agent-native.com/visual-edit`, keep it visible while tools register, and read the title before work. Never enter, copy, or request - Network accessline 75
`https://design.agent-native.com/visual-edit`, keep it visible while tools - Network accessline 78
other pages that block tools behind sign-in still require a signed-in tab. If the browser exposes CDP permissions, grant `local-network-access` to `https://design.agent-native.com` before calling page tools; otherwise use the - Network accessline 79
If the browser exposes CDP permissions, grant `local-network-access` to `https://design.agent-native.com` before calling page tools; otherwise use the page's Connect button and let the browser's permission prompt complete. - Network accessline 80
`https://design.agent-native.com` before calling page tools; otherwise use the - Network accessline 99
} const result = await an.call("open-visual-edit", { devServerUrl: "http://localhost:5173", - Network accessline 100
const result = await an.call("open-visual-edit", { devServerUrl: "http://localhost:5173", paths: ["/"], - Network accessline 101
devServerUrl: "http://localhost:5173", - Network accessline 175
- Browser hosts can use page-local `get-visual-edit-prompt`. - The `open-visual-edit` action is owned by Design. From another app, use the hosted MCP server at `https://design.agent-native.com/mcp` or the page's - Network accessline 176
- The `open-visual-edit` action is owned by Design. From another app, use the hosted MCP server at `https://design.agent-native.com/mcp` or the page's WebMCP helper, not `pnpm action` in the target app. The page path works - Network accessline 177
hosted MCP server at `https://design.agent-native.com/mcp` or the page's - Network accessline 198
path/query for another state; preserve the order of named or numbered flows. Shorthand like `localhost:1234/onboarding/1` means `http://localhost:1234/onboarding/1`. - Network accessline 200
`http://localhost:1234/onboarding/1`. - Network accessline 279
[Credential-related source omitted] - Network accessline 280
[Credential-related source omitted] - Access to secretsline 280
[Credential-related source omitted] - Network accessline 291
npx @agent-native/core@latest design connect --url http://localhost:5173 --root . --json - Network accessline 303
pnpm dev:cli design connect --url http://localhost:5173 --root templates/<app> --json - Network accessline 315
[Credential-related source omitted] - Network accessline 318
(Equivalently, pass `--bridge-token <token>`.) This starts a detached bridge on `http://127.0.0.1:7331`, adopts the server-minted token — so bridge and row - Network accessline 319
`http://127.0.0.1:7331`, adopts the server-minted token — so bridge and row - Network accessline 326
curl http://127.0.0.1:7331/health - Network accessline 371
{ "title": "Docs homepage visual edit", "devServerUrl": "http://localhost:5173", - Network accessline 372
"title": "Docs homepage visual edit", "devServerUrl": "http://localhost:5173", "bridgeUrl": "http://127.0.0.1:7331", - Network accessline 373
"devServerUrl": "http://localhost:5173", - Network accessline 374
"bridgeUrl": "http://127.0.0.1:7331", - Network accessline 394
{ "title": "Tasks responsive visual edit", "devServerUrl": "http://localhost:5173", - Network accessline 395
"title": "Tasks responsive visual edit", "devServerUrl": "http://localhost:5173", "bridgeUrl": "http://127.0.0.1:7331", - Network accessline 396
"devServerUrl": "http://localhost:5173", - Network accessline 397
"bridgeUrl": "http://127.0.0.1:7331", - Network accessline 431
"designId": "<existing-design-id>", "connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:5173", - Network accessline 432
"connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:5173", "paths": ["/settings", "/team"], - Network accessline 433
"devServerUrl": "http://localhost:5173", - Network accessline 447
"designId": "<existing-design-id>", "connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:1234", - Network accessline 448
"connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:1234", "routes": [ - Network accessline 449
"devServerUrl": "http://localhost:1234", - Network accessline 470
The fallback still targets `https://design.agent-native.com`; only the app and
skills/visual-plans/SKILL.md
- Network accessline 509
separate slash commands. Pass `--no-connect` to register the connector without authenticating, then run `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 510
authenticating, then run `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` whenever you are ready, or choose a narrower `--client`. Auth and MCP tool - Network accessline 511
`npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 535
Hosted default: connect `https://plan.agent-native.com/mcp`. Do
skills/visual-plans/references/connection.md
skills/visual-plans/references/local-files.md
- Network accessline 76
is `false`, verify used the offline lint because the app URL was remote or the local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on - Network accessline 77
local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the - Network accessline 78
`--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the concrete validation error.
skills/visual-recap/SKILL.md
- Network accessline 264
A recap lives only in the database of the MCP that created it. A separately running local dev server (e.g. `http://localhost:8081`) has its OWN database and - Network accessline 265
running local dev server (e.g. `http://localhost:8081`) has its OWN database and - Network accessline 273
`visualUrl` in the returned `plan.mdx` frontmatter, else `url`/`path` resolved against the MCP server's own origin (for the hosted MCP that is `https://plan.agent-native.com`). This always points at the database that has - Network accessline 274
resolved against the MCP server's own origin (for the hosted MCP that is `https://plan.agent-native.com`). This always points at the database that has the plan. - Network accessline 275
`https://plan.agent-native.com`). This always points at the database that has - Network accessline 277
2. Use a `localhost`/dev origin ONLY when the recap was created through a Plan MCP bound to that same origin — i.e. that MCP's url is `http://localhost:<port>/mcp`. Creating through the hosted MCP - Network accessline 278
MCP bound to that same origin — i.e. that MCP's url is `http://localhost:<port>/mcp`. Creating through the hosted MCP and linking to localhost is the exact mismatch that 404s. - Network accessline 279
`http://localhost:<port>/mcp`. Creating through the hosted MCP - Network accessline 280
and linking to localhost is the exact mismatch that 404s. 3. If only a plan id is available, build the MCP origin's absolute URL (hosted: `https://plan.agent-native.com/plans/<id>`) and say it was inferred. - Network accessline 282
(hosted: `https://plan.agent-native.com/plans/<id>`) and say it was inferred. - Network accessline 285
hosted MCP, say so plainly: the local dev server cannot see it. To view a recap on localhost (e.g. to exercise un-deployed local renderer changes), they must connect a LOCAL Plan MCP (`http://localhost:<port>/mcp`) and - Network accessline 286
on localhost (e.g. to exercise un-deployed local renderer changes), they must connect a LOCAL Plan MCP (`http://localhost:<port>/mcp`) and re-create the recap through it so it lands in the local database; offer to do - Network accessline 287
connect a LOCAL Plan MCP (`http://localhost:<port>/mcp`) and
skills/visual-recap/references/connection.md
skills/visual-recap/references/local-files.md
- Network accessline 76
is `false`, verify used the offline lint because the app URL was remote or the local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on - Network accessline 77
local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the - Network accessline 78
`--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the concrete validation error.
templates/analytics/.agents/skills/actions/SKILL.md
templates/analytics/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/analytics/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/analytics/.agents/skills/analysis-workspace/SKILL.md
- Network accessline 243
and example paths. 2. `provider-api-docs provider=<id> url=<docs-url>` — fetch any public docs or OpenAPI spec URL. Works for any `https://` URL. - Network accessline 244
2. `provider-api-docs provider=<id> url=<docs-url>` — fetch any public docs or OpenAPI spec URL. Works for any `https://` URL. 3. `provider-api-request` — call the endpoint directly. - Network accessline 245
OpenAPI spec URL. Works for any `https://` URL. - Network accessline 258
id="my-internal-api" label="My Internal API" baseUrl="https://api.mycompany.com" - Network accessline 259
label="My Internal API" baseUrl="https://api.mycompany.com" auth={ type: "bearer", credentialKey: "MY_API_TOKEN" } - Network accessline 260
baseUrl="https://api.mycompany.com" - Network accessline 262
docsUrls=["https://docs.mycompany.com/api"]
templates/analytics/.agents/skills/dashboard-management/SKILL.md
templates/analytics/.agents/skills/dashboard-ops/SKILL.md
templates/analytics/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/analytics/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/analytics/.agents/skills/onboarding-journeys/SKILL.md
- Network accessline 260
Authenticate to the deployed app, never a local database: run `npx -y @agent-native/core@latest connect https://analytics.agent-native.com --client codex` - Network accessline 261
`npx -y @agent-native/core@latest connect https://analytics.agent-native.com --client codex` - Access to secretsline 286
[Credential-related source omitted]
templates/analytics/.agents/skills/prometheus/SKILL.md
- Network accessline 14
| Env Var | Required | Notes | | ------------------------- | -------- | --------------------------------------------------------------------------- | | - Network accessline 15
| ------------------------- | -------- | --------------------------------------------------------------------------- | | `PROMETHEUS_URL` | yes | Base URL, no trailing slash, e.g. `https://prometheus.example.com`. | | - Network accessline 16
| `PROMETHEUS_URL` | yes | Base URL, no trailing slash, e.g. `https://prometheus.example.com`. | - Network accessline 73
brew install node_exporter prometheus brew services start node_exporter # metrics at http://localhost:9100/metrics - Network accessline 74
brew services start node_exporter # metrics at http://localhost:9100/metrics - Network accessline 90
brew services restart prometheus # Prometheus UI at http://localhost:9090 - Network accessline 93
Paste `http://localhost:9090` as the Prometheus URL in Data Sources, then use
templates/analytics/.agents/skills/provider-api/SKILL.md
- Network accessline 58
[Credential-related source omitted] - Access to secretsline 58
[Credential-related source omitted] - Network accessline 59
in the `clay-api-key` header. Never pass the key in action arguments. - Provider requests are restricted to the exact `https://api.clay.com` origin, with `/public/v0` as the default base path. Read the registered official docs - Network accessline 60
- Provider requests are restricted to the exact `https://api.clay.com` origin,
templates/analytics/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/analytics/.builder/skills/apollo/SKILL.md
- Access to secretsline 19
[Credential-related source omitted]
templates/analytics/.builder/skills/bigquery/SKILL.md
- Network accessline 74
| Cryptic Column | Meaning | Example Values | | -------------- | ------------------------ | ------------------------------------------------------------ | | `SUOHFYGIOG` | Blo - Network accessline 75
| -------------- | ------------------------ | ------------------------------------------------------------ | | `SUOHFYGIOG` | Blog URL | `https://www.example.com/blog/sample-post` | | `H5YIATNDT5` | Aut - Network accessline 76
| `SUOHFYGIOG` | Blog URL | `https://www.example.com/blog/sample-post` | - Network accessline 85
**Deduplication**: Table has duplicates (http:// vs https://). Always deduplicate: `REGEXP_EXTRACT(SUOHFYGIOG, r'/blog/([^/?#]+)')` with ROW_NUMBER or DISTINCT.
templates/analytics/.builder/skills/charts/SKILL.md
- Network accessline 63
**Example Corp Product Usage — Last 30 Days**  - Network accessline 64
 - Network accessline 70
When using QuickChart (`https://quickchart.io/chart/create`) with Chart.js v2:
templates/analytics/.builder/skills/dataforseo/SKILL.md
- Network accessline 12
- **Base URL**: `https://api.dataforseo.com/v3`
templates/analytics/.builder/skills/gong/SKILL.md
templates/analytics/.builder/skills/grafana/SKILL.md
templates/analytics/.builder/skills/hubspot/SKILL.md
templates/analytics/.builder/skills/jira/SKILL.md
- Network accessline 12
- **Base URL**: `$JIRA_BASE_URL` (e.g. `https://yourorg.atlassian.net`) - Access to secretsline 13
[Credential-related source omitted] - Access to secretsline 14
[Credential-related source omitted] - Network accessline 63
- **IMPORTANT**: The old `/rest/api/3/search` endpoint was removed by Atlassian (returns 410). Must use `/rest/api/3/search/jql` instead. - JQL for duplicate detection: `project = X AND summary ~ "keyword" ORDER BY created DESC` - Always us - Network accessline 64
- JQL for duplicate detection: `project = X AND summary ~ "keyword" ORDER BY created DESC` - Always use markdown links for tickets: `[ENG-1234](https://yourorg.atlassian.net/browse/ENG-1234)` - `getStatuses` with a projectKey returns an arr - Network accessline 65
- Always use markdown links for tickets: `[ENG-1234](https://yourorg.atlassian.net/browse/ENG-1234)`
templates/analytics/.builder/skills/notion/SKILL.md
templates/analytics/.builder/skills/provider-api/SKILL.md
- Network accessline 23
[Credential-related source omitted] - Access to secretsline 23
[Credential-related source omitted] - Network accessline 24
`clay-api-key` header and restricts requests to the exact `https://api.clay.com` origin, with `/public/v0` as the default base path. Use `provider-api-docs` for the registered official docs and OpenAPI spec on - Network accessline 25
`https://api.clay.com` origin, with `/public/v0` as the default base path. Use
templates/analytics/.builder/skills/sentry/SKILL.md
templates/analytics/.builder/skills/slack/SKILL.md
templates/analytics/.builder/skills/stripe/SKILL.md
- Network accessline 10
- **Base URL**: `https://api.stripe.com`
templates/assets/.agents/skills/actions/SKILL.md
templates/assets/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/assets/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/assets/.agents/skills/agent-engines/SKILL.md
- Access to secretsline 47
[Credential-related source omitted] - Access to secretsline 108
[Credential-related source omitted] - Access to secretsline 109
[Credential-related source omitted] - Access to secretsline 110
[Credential-related source omitted] - Access to secretsline 111
[Credential-related source omitted] - Access to secretsline 112
[Credential-related source omitted] - Access to secretsline 113
[Credential-related source omitted] - Access to secretsline 114
[Credential-related source omitted] - Access to secretsline 115
[Credential-related source omitted] - Network accessline 149
Any `vendor/model` string from [openrouter.ai/models](https://openrouter.ai/models) works — the `supportedModels` list in the registry is a UI hint, not an allow-list. - Network accessline 155
[Credential-related source omitted] - Access to secretsline 155
[Credential-related source omitted] - Network accessline 156
appName: "My App", appUrl: "https://myapp.example", }); - Network accessline 157
appUrl: "https://myapp.example", - Access to secretsline 188
[Credential-related source omitted] - Access to secretsline 225
[Credential-related source omitted] - Access to secretsline 226
[Credential-related source omitted] - Access to secretsline 227
[Credential-related source omitted] - Access to secretsline 228
[Credential-related source omitted] - Access to secretsline 229
[Credential-related source omitted] - Access to secretsline 230
[Credential-related source omitted] - Access to secretsline 231
[Credential-related source omitted]
templates/assets/.agents/skills/asset-generation/SKILL.md
- Network accessline 148
- Hosted default: connect `https://assets.agent-native.com/_agent-native/mcp`.
templates/assets/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/assets/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/assets/.agents/skills/notifications/SKILL.md
templates/assets/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/brain/.agents/skills/actions/SKILL.md
templates/brain/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/brain/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/brain/.agents/skills/brain/RUNBOOK.md
- Access to secretsline 54
[Credential-related source omitted] - Network accessline 342
"occurredAt": "2026-05-15T15:00:00.000Z", "transcript": "We decided to keep annual pricing because...", "sourceUrl": "https://example.com/share/meeting-123", - Network accessline 343
"transcript": "We decided to keep annual pricing because...", "sourceUrl": "https://example.com/share/meeting-123", "tags": ["pricing", "product"], - Network accessline 344
"sourceUrl": "https://example.com/share/meeting-123", - Network accessline 370
"content": "Agent-Native is ...", "capturedAt": "2026-07-30T12:00:00.000Z", "sourceUrl": "https://docs.example.com/what-is-agent-native", - Network accessline 371
"capturedAt": "2026-07-30T12:00:00.000Z", "sourceUrl": "https://docs.example.com/what-is-agent-native", "tags": ["faq", "agent-native"], - Network accessline 372
"sourceUrl": "https://docs.example.com/what-is-agent-native",
templates/brain/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/brain/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/brain/.agents/skills/ingestion-and-connectors/SKILL.md
- Access to secretsline 49
[Credential-related source omitted]
templates/brain/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/calendar/.agents/skills/actions/SKILL.md
templates/calendar/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/calendar/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/calendar/.agents/skills/event-management/SKILL.md
- Network accessline 225
Use `--attachments '[{"fileUrl":"https://drive.google.com/...","title":"Agenda"}]'` to attach Drive files, HTTPS file links, or files uploaded through the app's file upload storage. Google Calendar supports up to 25 attachments per event. - Network accessline 314
pnpm action update-event --id google-event-id --reminders '[{"method":"popup","minutes":10},{"method":"email","minutes":1440}]' --colorId 9 --attachments '[{"fileUrl":"https://drive.google.com/...","title":"Agenda"}]' - Network accessline 316
--reminders '[{"method":"popup","minutes":10},{"method":"email","minutes":1440}]' \ --colorId 9 \ --attachments '[{"fileUrl":"https://drive.google.com/...","title":"Agenda"}]' - Network accessline 318
--attachments '[{"fileUrl":"https://drive.google.com/...","title":"Agenda"}]' - Network accessline 550
], "conferenceData": { ... }, "hangoutLink": "https://meet.google.com/...", - Network accessline 551
"conferenceData": { ... }, "hangoutLink": "https://meet.google.com/...", "status": "confirmed", - Network accessline 552
"hangoutLink": "https://meet.google.com/...",
templates/calendar/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/calendar/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/calendar/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/chat/.agents/skills/actions/SKILL.md
templates/chat/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/chat/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/chat/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/clips/.agents/skills/actions/SKILL.md
templates/clips/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/clips/.agents/skills/ai-video-tools/SKILL.md
- Access to secretsline 14
[Credential-related source omitted] - Access to secretsline 52
[Credential-related source omitted] - Network accessline 130
1. **Native (highest priority).** The browser's Web Speech API, desktop local Whisper/macOS SFSpeech when available, and desktop Web Speech fallback on non-mac run during recording and save results via `save-browser-transcript`. This gives - Network accessline 132
[Credential-related source omitted] - Access to secretsline 132
[Credential-related source omitted] - Access to secretsline 206
[Credential-related source omitted] - Network accessline 208
description: "Fast speech-to-text fallback via Groq. Builder Gemini Flash-Lite is preferred when connected; Groq is used only when Builder/native transcription is unavailable.", docsUrl: "https://console.groq.com/keys", - Network accessline 209
"Fast speech-to-text fallback via Groq. Builder Gemini Flash-Lite is preferred when connected; Groq is used only when Builder/native transcription is unavailable.", docsUrl: "https://console.groq.com/keys", scope: "user", - Network accessline 210
docsUrl: "https://console.groq.com/keys",
templates/clips/.agents/skills/dictate/SKILL.md
- Access to secretsline 60
[Credential-related source omitted]
templates/clips/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/clips/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/clips/.agents/skills/meetings/SKILL.md
- Access to secretsline 129
[Credential-related source omitted]
templates/clips/.agents/skills/recording/SKILL.md
templates/clips/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/clips/.agents/skills/video-sharing/SKILL.md
- Access to secretsline 204
[Credential-related source omitted]
templates/content/.agents/skills/actions/SKILL.md
templates/content/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/content/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/content/.agents/skills/document-editing/SKILL.md
- Network accessline 110
pnpm action import-content --dryRun true --parentId abc123 --files '[{"name":"guide.md","text":"# Guide\n..."},{"name":"logo.png"}]' pnpm action import-content --dryRun false --parentId abc123 --idempotencyKey guide-1 --files '[{"name":"gui - Network accessline 111
pnpm action import-content --dryRun false --parentId abc123 --idempotencyKey guide-1 --files '[{"name":"guide.md","text":"..."},{"name":"logo.png","url":"https://..."}]' - Network accessline 296
Pass `quotedText` exactly as the page reads: text from one block, without Markdown syntax such as `**`, backslash escapes, or a link's URL (quote `docs`, not `[docs](https://…)`). `add-comment` refuses a quote the saved page or tab doesn't - Network accessline 347
pnpm action create-content-webhook --databaseId db123 --url https://example.com/hook \
templates/content/.agents/skills/document-editing/references/document-behavior.md
- Network accessline 67
Documents support image blocks as markdown images: ``.
templates/content/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/content/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/content/.agents/skills/notion-integration/SKILL.md
- Access to secretsline 155
[Credential-related source omitted]
templates/content/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/crm/.agents/skills/actions/SKILL.md
templates/crm/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/crm/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/crm/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/crm/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/crm/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/design/.agents/skills/actions/SKILL.md
templates/design/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/design/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/design/.agents/skills/design-editor-performance/references/measuring.md
templates/design/.agents/skills/design-generation/SKILL.md
- Network accessline 618
<!-- Tailwind v4 browser runtime --> <script src="https://cdn.jsdelivr.net/npm/@tailwindcss/browser@4"></script> - Network accessline 619
<script src="https://cdn.jsdelivr.net/npm/@tailwindcss/browser@4"></script> - Network accessline 621
<!-- Alpine.js — MUST be deferred --> <script defer src="https://cdn.jsdelivr.net/npm/alpinejs@3.15.11/dist/cdn.min.js"></script> - Network accessline 622
<script defer src="https://cdn.jsdelivr.net/npm/alpinejs@3.15.11/dist/cdn.min.js"></script> - Network accessline 624
<!-- Google Fonts — pick distinctive fonts --> <link rel="preconnect" href="https://fonts.googleapis.com"> - Network accessline 625
<link rel="preconnect" href="https://fonts.googleapis.com"> - Network accessline 626
<link rel="preconnect" href="https://fonts.gstatic.com" crossorigin> - Network accessline 627
<link href="https://fonts.googleapis.com/css2?family=...&display=swap" rel="stylesheet"> - Network accessline 941
External links (`https://…`) are allowed — the editor opens them in a new tab.
templates/design/.agents/skills/design-systems/SKILL.md
- Network accessline 212
pnpm action import-from-url --url "https://acme.com" - Network accessline 225
pnpm action index-design-system-with-builder --githubSources '[{"repoUrl":"https://github.com/acme/ui","ref":"main","include":["src/styles","design.md"]}]' - Network accessline 357
pnpm action import-figma-frame --figmaUrl "https://www.figma.com/design/<fileKey>/<name>?node-id=<id>" - Access to secretsline 427
[Credential-related source omitted] - Network accessline 458
# No nodeId: lists pages and top-level frames (like get_metadata with no node id) pnpm action get-figma-design-context --figmaUrl "https://www.figma.com/design/<fileKey>/<name>" - Network accessline 459
pnpm action get-figma-design-context --figmaUrl "https://www.figma.com/design/<fileKey>/<name>" - Network accessline 461
pnpm action get-figma-design-context --figmaUrl "https://www.figma.com/design/<fileKey>/<name>?node-id=<id>" - Access to secretsline 500
[Credential-related source omitted] - Network accessline 505
pnpm action analyze-brand-assets \ --websiteUrl "https://acme.com" \ - Network accessline 506
--websiteUrl "https://acme.com" \ - Network accessline 582
<link href="https://fonts.googleapis.com/css2?family=HEADING_FONT:wght@400;700;900&family=BODY_FONT:wght@300;400;600&display=swap" rel="stylesheet">
templates/design/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/design/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/design/.agents/skills/hubspot-email/SKILL.md
- Network accessline 52
`unsubscribe_link_all` URLs opt out of this email type or all email; `view_as_page_url` opens the web version. Use the [official HubL variable reference](https://developers.hubspot.com/docs/cms/reference/hubl/variables). - Network accessline 53
`view_as_page_url` opens the web version. Use the [official HubL variable reference](https://developers.hubspot.com/docs/cms/reference/hubl/variables). Keep footer text and a nonempty unsubscribe label visible outside the preheader, - Network accessline 54
[official HubL variable reference](https://developers.hubspot.com/docs/cms/reference/hubl/variables). - Network accessline 62
support personalization filters in email; CMS/blog examples do not establish email support. The [official function reference](https://developers.hubspot.com/docs/cms/reference/hubl/functions#personalization-token) - Network accessline 63
email support. The [official function reference](https://developers.hubspot.com/docs/cms/reference/hubl/functions#personalization-token) documents the default argument, and the - Network accessline 64
[official function reference](https://developers.hubspot.com/docs/cms/reference/hubl/functions#personalization-token) - Network accessline 65
documents the default argument, and the [filter reference](https://developers.hubspot.com/docs/cms/reference/hubl/filters) excludes personalization filters in emails. Dynamic CTA defaults must be - Network accessline 66
[filter reference](https://developers.hubspot.com/docs/cms/reference/hubl/filters)
templates/design/.agents/skills/journey-storyboards/SKILL.md
- Network accessline 119
- `imageUrl` must be `https://`. `data:` URLs, other schemes and embedded credentials are rejected.
templates/design/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/design/.agents/skills/visual-edit/SKILL.md
- Network accessline 18
The editor is hosted at `https://design.agent-native.com`. Never start local - Network accessline 21
See the [Visual Edit guide](https://github.com/BuilderIO/agent-native/blob/main/skills/visual-edit/README.md) - Network accessline 57
- Inline-browser hosts should open `https://design.agent-native.com/visual-edit` - Network accessline 73
JavaScript evaluator. Prefer the host's inline browser; use external Chrome only when requested or unavailable. Open `https://design.agent-native.com/visual-edit`, keep it visible while tools - Network accessline 74
only when requested or unavailable. Open `https://design.agent-native.com/visual-edit`, keep it visible while tools register, and read the title before work. Never enter, copy, or request - Network accessline 75
`https://design.agent-native.com/visual-edit`, keep it visible while tools - Network accessline 78
other pages that block tools behind sign-in still require a signed-in tab. If the browser exposes CDP permissions, grant `local-network-access` to `https://design.agent-native.com` before calling page tools; otherwise use the - Network accessline 79
If the browser exposes CDP permissions, grant `local-network-access` to `https://design.agent-native.com` before calling page tools; otherwise use the page's Connect button and let the browser's permission prompt complete. - Network accessline 80
`https://design.agent-native.com` before calling page tools; otherwise use the - Network accessline 99
} const result = await an.call("open-visual-edit", { devServerUrl: "http://localhost:5173", - Network accessline 100
const result = await an.call("open-visual-edit", { devServerUrl: "http://localhost:5173", paths: ["/"], - Network accessline 101
devServerUrl: "http://localhost:5173", - Network accessline 175
- Browser hosts can use page-local `get-visual-edit-prompt`. - The `open-visual-edit` action is owned by Design. From another app, use the hosted MCP server at `https://design.agent-native.com/mcp` or the page's - Network accessline 176
- The `open-visual-edit` action is owned by Design. From another app, use the hosted MCP server at `https://design.agent-native.com/mcp` or the page's WebMCP helper, not `pnpm action` in the target app. The page path works - Network accessline 177
hosted MCP server at `https://design.agent-native.com/mcp` or the page's - Network accessline 198
path/query for another state; preserve the order of named or numbered flows. Shorthand like `localhost:1234/onboarding/1` means `http://localhost:1234/onboarding/1`. - Network accessline 200
`http://localhost:1234/onboarding/1`. - Network accessline 279
[Credential-related source omitted] - Network accessline 280
[Credential-related source omitted] - Access to secretsline 280
[Credential-related source omitted] - Network accessline 291
npx @agent-native/core@latest design connect --url http://localhost:5173 --root . --json - Network accessline 303
pnpm dev:cli design connect --url http://localhost:5173 --root templates/<app> --json - Network accessline 315
[Credential-related source omitted] - Network accessline 318
(Equivalently, pass `--bridge-token <token>`.) This starts a detached bridge on `http://127.0.0.1:7331`, adopts the server-minted token — so bridge and row - Network accessline 319
`http://127.0.0.1:7331`, adopts the server-minted token — so bridge and row - Network accessline 326
curl http://127.0.0.1:7331/health - Network accessline 371
{ "title": "Docs homepage visual edit", "devServerUrl": "http://localhost:5173", - Network accessline 372
"title": "Docs homepage visual edit", "devServerUrl": "http://localhost:5173", "bridgeUrl": "http://127.0.0.1:7331", - Network accessline 373
"devServerUrl": "http://localhost:5173", - Network accessline 374
"bridgeUrl": "http://127.0.0.1:7331", - Network accessline 394
{ "title": "Tasks responsive visual edit", "devServerUrl": "http://localhost:5173", - Network accessline 395
"title": "Tasks responsive visual edit", "devServerUrl": "http://localhost:5173", "bridgeUrl": "http://127.0.0.1:7331", - Network accessline 396
"devServerUrl": "http://localhost:5173", - Network accessline 397
"bridgeUrl": "http://127.0.0.1:7331", - Network accessline 431
"designId": "<existing-design-id>", "connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:5173", - Network accessline 432
"connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:5173", "paths": ["/settings", "/team"], - Network accessline 433
"devServerUrl": "http://localhost:5173", - Network accessline 447
"designId": "<existing-design-id>", "connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:1234", - Network accessline 448
"connectionId": "<existing-connection-id>", "devServerUrl": "http://localhost:1234", "routes": [ - Network accessline 449
"devServerUrl": "http://localhost:1234", - Network accessline 470
The fallback still targets `https://design.agent-native.com`; only the app and
templates/dispatch/.agents/skills/a2a-protocol/SKILL.md
- Network accessline 65
"name": "Analytics", "description": "Queries analytics data across providers", "url": "https://analytics.example.com", - Network accessline 66
"description": "Queries analytics data across providers", "url": "https://analytics.example.com", "color": "#6B7280", - Network accessline 67
"url": "https://analytics.example.com", - Network accessline 68
"color": "#6B7280", "cardUrl": "https://analytics.example.com/.well-known/agent-card.json", "auth": { - Network accessline 69
"cardUrl": "https://analytics.example.com/.well-known/agent-card.json", - Network accessline 102
The last one matters for template apps with no Code tab: "connect me to https://analytics.example.com" is a complete instruction. `remote-agents/` is - Network accessline 103
https://analytics.example.com" is a complete instruction. `remote-agents/` is - Network accessline 107
First-party templates are seeded automatically and overlaid on top of manifests, so `Mail`, `Calendar`, and friends appear without registration. In dev they resolve to `http://localhost:<devPort>`; a manifest pointing at localhost is - Network accessline 108
so `Mail`, `Calendar`, and friends appear without registration. In dev they resolve to `http://localhost:<devPort>`; a manifest pointing at localhost is ignored in production. - Network accessline 109
resolve to `http://localhost:<devPort>`; a manifest pointing at localhost is - Network accessline 210
`.../agents/{agent}/endpoint/protocols/a2a/agentCard/v1.0` when available, and pass `protocolVersion` only when the card omits it. Foundry v1 does not provide SSE streaming. See the [Foundry hosted agent overview](https://learn.microsoft.co - Network accessline 211
and pass `protocolVersion` only when the card omits it. Foundry v1 does not provide SSE streaming. See the [Foundry hosted agent overview](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) and [A2A endpoint guidance](https:// - Network accessline 212
provide SSE streaming. See the [Foundry hosted agent overview](https://learn.microsoft.com/en-us/azure/foundry/agents/overview) - Network accessline 213
and [A2A endpoint guidance](https://learn.microsoft.com/en-us/azure/foundry/agents/how-to/enable-agent-to-agent-endpoint). - Access to secretsline 237
[Credential-related source omitted] - Network accessline 291
const auth = await resolveA2ACallerAuth(); const answer = await callAgent( "https://analytics.example.com", - Network accessline 292
const answer = await callAgent( "https://analytics.example.com", "What were last week's signups?", - Network accessline 293
"https://analytics.example.com", - Network accessline 366
// The second argument is the bearer token itself, not a signing secret. const { apiKey, apiKeyFallbacks } = await resolveA2ACallerAuth(); const client = new A2AClient("https://analytics.example.com", apiKey, { - Network accessline 367
const { apiKey, apiKeyFallbacks } = await resolveA2ACallerAuth(); const client = new A2AClient("https://analytics.example.com", apiKey, { fallbackApiKeys: apiKeyFallbacks, - Network accessline 368
const client = new A2AClient("https://analytics.example.com", apiKey, {
templates/dispatch/.agents/skills/actions/SKILL.md
templates/dispatch/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/dispatch/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/dispatch/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/dispatch/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/dispatch/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/dispatch/.agents/skills/turn-into-app/references/attachments.md
templates/dispatch/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
templates/dispatch/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
templates/dispatch/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
templates/dispatch/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
templates/dispatch/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
templates/factory/.agents/skills/actions/SKILL.md
templates/factory/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/factory/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/factory/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/factory/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/factory/.agents/skills/review-latest-feedback/SKILL.md
- Network accessline 102
- `thanks - the <https://github.com/BuilderIO/agent-native/pull/1234|fix> so the export keeps your column order is merged and will be on beta shortly.` - Network accessline 103
- `thanks - once this <https://github.com/BuilderIO/agent-native/pull/1234|fix> merges, the export will keep your column order; it's in review now.` - Network accessline 110
thanks - the <https://github.com/BuilderIO/agent-native/pull/1234|fix> so the export keeps your column order is merged and will be on beta shortly.
templates/factory/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/factory/.agents/skills/turn-into-app/references/attachments.md
templates/factory/.agents/skills/turn-into-app/references/fresh-project.md
- Network accessline 69
instructions in the Project instructions or knowledge files. - Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the - Network accessline 70
- Add and authenticate the Agent-Native Dispatch MCP connector (`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the workspace's Builder project through the Builder Projects API, so no separate - Network accessline 71
(`https://dispatch.agent-native.com/mcp`, OAuth). It reuses or provisions the
templates/factory/.agents/skills/turn-into-app/references/local-run-and-deploy.md
- Network accessline 52
defaults only. Precedence, supported modes, and the line between committed config and deployment secrets: https://www.agent-native.com/docs/agent-native-config - Network accessline 54
https://www.agent-native.com/docs/agent-native-config - Access to secretsline 71
[Credential-related source omitted] - Access to secretsline 72
[Credential-related source omitted] - Network accessline 85
mkdir -p .tmp && (nohup pnpm exec agent-native dev --port <port> > .tmp/dev.log 2>&1 & echo $! > .tmp/dev.pid) wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null - - Network accessline 86
wait200() { while t=$((end-SECONDS)); [ $t -gt 0 ]; do [ "$(curl -sL --connect-timeout 2 --max-time $((t<10?t:10)) -o /dev/null -w '%{http_code}' "$1")" = 200 ] && return 0; sleep 1; done; echo "no 200 from $1 before the three-minute deadli - Network accessline 87
end=$((SECONDS+180)); wait200 http://localhost:<port>/ && wait200 http://localhost:<port>/<route> - Network accessline 92
assignment, and each request is capped by the time left; run that last line again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. Th - Network accessline 93
again after every restart for a fresh clock. A nonzero exit means the server or the route never answered: read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the server can answer; - Network accessline 94
read `.tmp/dev.log` and fix that before any screenshot. The log prints `Local: http://localhost:<port>/` before the - Network accessline 137
`pnpm build`, then the provider's own deploy command. Local PGlite storage is not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - Network accessline 138
not production storage. Guide: https://www.agent-native.com/docs/deploy-an-app - **App inside a workspace:** deploy the workspace from its root, which needs - Network accessline 139
https://www.agent-native.com/docs/deploy-an-app - Network accessline 148
For Vercel, run `npx @agent-native/core@latest deploy --preset vercel`, then `vercel deploy --prebuilt`. Guide: https://www.agent-native.com/docs/workspace-deployment - Network accessline 150
https://www.agent-native.com/docs/workspace-deployment
templates/factory/.agents/skills/turn-into-app/references/review-loop.md
- Network accessline 39
node .tmp/ui-review/shoot.mjs http://localhost:<port>/ p1 'button:has-text("<agent verb>")' /<route>
templates/factory/.agents/skills/turn-into-app/references/spreadsheet-source.md
- Network accessline 50
host, or, in Dispatch or Analytics, the `google_drive` provider through `provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the - Network accessline 51
`provider-api-catalog`, `provider-api-docs`, and `provider-api-request` with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the provider's base is Drive v3). The chat scaffold's `provider-api-request` is - Network accessline 52
with a full `https://sheets.googleapis.com/v4/spreadsheets/<id>` URL (the
templates/factory/.agents/skills/turn-into-app/references/where-it-runs.md
- Network accessline 78
1. Connect Dispatch. If it is not configured, run `npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 79
`npx @agent-native/core@latest connect https://dispatch.agent-native.com` - Network accessline 80
with `--client` set to this host. Without Node.js 22.22+, don't make the person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each - Network accessline 81
person install it just to pick Builder: send them to `https://dispatch.agent-native.com/mcp/connect`, which has steps for each host. If it is configured but signed out, sign in from the host's MCP - Network accessline 82
`https://dispatch.agent-native.com/mcp/connect`, which has steps for each
templates/forms/.agents/skills/actions/SKILL.md
templates/forms/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/forms/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/forms/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/forms/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/forms/.agents/skills/form-publishing/SKILL.md
- Network accessline 42
The slug is auto-generated from the title + a short unique suffix: - Title: "Contact Form" -> Slug: `contact-form/a1b2c3` - Full URL: `https://yourapp.com/f/contact-form/a1b2c3` - Network accessline 44
- Full URL: `https://yourapp.com/f/contact-form/a1b2c3` - Access to secretsline 87
[Credential-related source omitted]
templates/forms/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/mail/.agents/skills/actions/SKILL.md
templates/mail/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/mail/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/mail/.agents/skills/email-drafts/SKILL.md
- Network accessline 88
- **Links:** `[click here](https://example.com)` — renders as a clickable hyperlink in the sent email - Network accessline 93
- **Code:** `` `inline code` `` or fenced code blocks - **Blockquotes:** `> quoted text` - **Bare URLs:** `https://example.com` auto-links - Network accessline 95
- **Bare URLs:** `https://example.com` auto-links - Network accessline 203
curl --fail-with-body --request PUT --header '<Authorization header>' --data-binary @/absolute/path/to/report.pdf '<uploadUrl>'
templates/mail/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/mail/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/mail/.agents/skills/mail-interaction-parity/references/2026-09-12-initial-pass.md
- Access to secretsline 138
[Credential-related source omitted] - Network accessline 159
draft, SVG/text-node keyboard targets, Tab/Shift+Tab cycling, and conditional combobox `aria-controls` references. - The local Mail app started at `http://localhost:8080/inbox`. Selecting - Network accessline 160
combobox `aria-controls` references. - The local Mail app started at `http://localhost:8080/inbox`. Selecting **Connect Google** reached - Network accessline 161
- The local Mail app started at `http://localhost:8080/inbox`. Selecting - Access to secretsline 165
[Credential-related source omitted]
templates/mail/.agents/skills/mail-interaction-parity/references/interaction-matrix.md
- Network accessline 25
Captured evidence for the post-fix head `7d080b8` is available in the [dark visual recap image](https://plan.agent-native.com/_agent-native/recap-image/e56aa10a91f59d04e12cb7e2e816fc8f0700614d7ef240fe1fba40a2f6162edb.png?v=34988715504-1) - Network accessline 26
[dark visual recap image](https://plan.agent-native.com/_agent-native/recap-image/e56aa10a91f59d04e12cb7e2e816fc8f0700614d7ef240fe1fba40a2f6162edb.png?v=34988715504-1) - Network accessline 27
and [light visual recap image](https://plan.agent-native.com/_agent-native/recap-image/af44730d61a368149a97eb899203d847433e574878dd979ff1610cc2b2b1205f.png?v=34988715504-1). - Network accessline 118
focus/scroll. Treat the documented iOS swipe-right-to-return and Android system Back outcomes as separate contracts; do not infer one from the other or from desktop Escape. Use the [Mobile Navigation](https://help.superhuman.com/hc/en-u - Network accessline 119
system Back outcomes as separate contracts; do not infer one from the other or from desktop Escape. Use the [Mobile Navigation](https://help.superhuman.com/hc/en-us/articles/46005719737357-Mobile-Navigation) guide as the reference, and - Network accessline 120
or from desktop Escape. Use the [Mobile Navigation](https://help.superhuman.com/hc/en-us/articles/46005719737357-Mobile-Navigation) - Network accessline 164
Mail source audit found contact/local matches start at two trimmed characters and automatic navigation to `/all?q=...` starts at three after 400 ms. Superhuman's official [Search in Seconds](https://help.superhuman.com/hc/en-us/articles - Network accessline 165
and automatic navigation to `/all?q=...` starts at three after 400 ms. Superhuman's official [Search in Seconds](https://help.superhuman.com/hc/en-us/articles/46005814266253-Search-in-Seconds) workflow says to type a query and press Ent - Network accessline 166
Superhuman's official [Search in Seconds](https://help.superhuman.com/hc/en-us/articles/46005814266253-Search-in-Seconds) - Network accessline 190
selecting an operator runs search or only edits the query. Record provider- and operator-specific support; do not infer that every Gmail operator is parsed identically. Use the official [Search](https://help.superhuman.com/hc/en-us/arti - Network accessline 191
and operator-specific support; do not infer that every Gmail operator is parsed identically. Use the official [Search](https://help.superhuman.com/hc/en-us/articles/46005672652301-Search) guide as reference. This planned comparison is n - Network accessline 192
parsed identically. Use the official [Search](https://help.superhuman.com/hc/en-us/articles/46005672652301-Search) - Network accessline 235
messages remain available without assuming the eviction order, and verify the per-Split cache limit. Search for a cached match and an offline miss, then reconnect and repeat. Apply the documented [Offline Access](https://help.superhuman - Network accessline 236
the per-Split cache limit. Search for a cached match and an offline miss, then reconnect and repeat. Apply the documented [Offline Access](https://help.superhuman.com/hc/en-us/articles/46005499629325-Offline-Access) eligibility and per- - Network accessline 237
reconnect and repeat. Apply the documented [Offline Access](https://help.superhuman.com/hc/en-us/articles/46005499629325-Offline-Access) - Network accessline 282
partition. On mobile, verify the visible Undo control, location, label, duration, replacement by a subsequent action, and tap result against the official [Undo](https://help.superhuman.com/hc/en-us/articles/46005666743309-Undo) - Network accessline 283
duration, replacement by a subsequent action, and tap result against the official [Undo](https://help.superhuman.com/hc/en-us/articles/46005666743309-Undo) reference. If the responsive Mail surface has no on-screen Undo, record the - Network accessline 284
official [Undo](https://help.superhuman.com/hc/en-us/articles/46005666743309-Undo) - Network accessline 285
reference. If the responsive Mail surface has no on-screen Undo, record the exact parity gap instead of treating the keyboard shortcut as equivalent. Official reference: Superhuman's [Undo](https://help.superhuman.com/hc/en-us/articles/ - Network accessline 286
exact parity gap instead of treating the keyboard shortcut as equivalent. Official reference: Superhuman's [Undo](https://help.superhuman.com/hc/en-us/articles/46005666743309-Undo) guide documents `Z` to undo the last action within 10 s - Network accessline 287
Official reference: Superhuman's [Undo](https://help.superhuman.com/hc/en-us/articles/46005666743309-Undo) - Network accessline 309
and the resulting visible bar. Verify the documented Android availability boundary rather than assuming iOS customization exists there. Use [Customizing Swipes and Triage Bar](https://help.superhuman.com/hc/en-us/articles/46005742942861 - Network accessline 310
boundary rather than assuming iOS customization exists there. Use [Customizing Swipes and Triage Bar](https://help.superhuman.com/hc/en-us/articles/46005742942861-Customizing-Swipes-and-Triage-Bar) as the platform reference. Mail source - Network accessline 311
[Customizing Swipes and Triage Bar](https://help.superhuman.com/hc/en-us/articles/46005742942861-Customizing-Swipes-and-Triage-Bar) - Network accessline 569
keyboard open/closed, selection handling, and whether formatting applies to selected text or the insertion point. Use the official [Formatting Text](https://help.superhuman.com/hc/en-us/articles/46005721681165-Formatting-Text) - Network accessline 570
selected text or the insertion point. Use the official [Formatting Text](https://help.superhuman.com/hc/en-us/articles/46005721681165-Formatting-Text) guide; it does not enumerate the full mobile Format menu, so do not assume - Network accessline 571
[Formatting Text](https://help.superhuman.com/hc/en-us/articles/46005721681165-Formatting-Text) - Network accessline 589
interpreting a missing suggestion: test desktop English, mobile, and a non-English compose locale, recording whether the feature is absent, disabled, or produces no match. The official [Autocomplete](https://help.superhuman.com/hc/en-us - Network accessline 590
non-English compose locale, recording whether the feature is absent, disabled, or produces no match. The official [Autocomplete](https://help.superhuman.com/hc/en-us/articles/46005685782669-Autocomplete) article documents phrase autocom - Network accessline 591
or produces no match. The official [Autocomplete](https://help.superhuman.com/hc/en-us/articles/46005685782669-Autocomplete) - Network accessline 640
long-press View attachment versus Save attachment. Include cancel/failure and return-to-thread state. Use the official [Attachments](https://help.superhuman.com/hc/en-us/articles/46005568142989-Attachments) - Network accessline 641
return-to-thread state. Use the official [Attachments](https://help.superhuman.com/hc/en-us/articles/46005568142989-Attachments) guide; picker paths may differ by platform. - Network accessline 642
[Attachments](https://help.superhuman.com/hc/en-us/articles/46005568142989-Attachments) - Network accessline 749
after send succeeds; if archive fails, preserve the sent state and do not retry the send. Mail now implements these paths locally, based on the [official Send + Mark Done guide](https://help.superhuman.com/hc/en-us/articles/474391346137 - Network accessline 750
retry the send. Mail now implements these paths locally, based on the [official Send + Mark Done guide](https://help.superhuman.com/hc/en-us/articles/47439134613773-Mark-Done). A 2026-09-14 paired palette replay confirmed Superhuman exp - Network accessline 751
[official Send + Mark Done guide](https://help.superhuman.com/hc/en-us/articles/47439134613773-Mark-Done). - Network accessline 779
reopen state, and that scheduling never falls through to Send. Record whether the mobile surface omits natural-language input as documented. Compare the official [Schedule Emails](https://help.superhuman.com/hc/en-us/articles/4720676385 - Network accessline 780
the mobile surface omits natural-language input as documented. Compare the official [Schedule Emails](https://help.superhuman.com/hc/en-us/articles/47206763851533-Switching-from-Notion-Mail-to-Superhuman-Mail) guide with the live behavi - Network accessline 781
official [Schedule Emails](https://help.superhuman.com/hc/en-us/articles/47206763851533-Switching-from-Notion-Mail-to-Superhuman-Mail) - Network accessline 931
account order after adding, switching, and reordering, and that the selected identity is consistent in the picker and compose From field. Use the official [Managing Accounts](https://help.superhuman.com/hc/en-us/articles/46005777934733- - Network accessline 932
identity is consistent in the picker and compose From field. Use the official [Managing Accounts](https://help.superhuman.com/hc/en-us/articles/46005777934733-Managing-Accounts) guide as reference. Compare Windows' documented account-sw - Network accessline 933
[Managing Accounts](https://help.superhuman.com/hc/en-us/articles/46005777934733-Managing-Accounts) - Network accessline 1032
and calendar; test Colemak's listed reply/navigation/snippet overrides. Do not infer native support for an unlisted international layout. The official [International keyboard shortcuts](https://help.superhuman.com/hc/en-us/articles/4600 - Network accessline 1033
infer native support for an unlisted international layout. The official [International keyboard shortcuts](https://help.superhuman.com/hc/en-us/articles/46005584339597-Shortcuts-for-International-Keyboards) article documents Shift+7 for - Network accessline 1034
[International keyboard shortcuts](https://help.superhuman.com/hc/en-us/articles/46005584339597-Shortcuts-for-International-Keyboards) - Network accessline 1066
the corresponding Mail controls and compare each tooltip with the action it triggers. Use the official [Keyboard shortcuts](https://help.superhuman.com/hc/en-us/articles/46005701270541-Keyboard-Shortcuts-in-Superhuman-Mail) - Network accessline 1067
triggers. Use the official [Keyboard shortcuts](https://help.superhuman.com/hc/en-us/articles/46005701270541-Keyboard-Shortcuts-in-Superhuman-Mail) guide as reference. Compare every displayed shortcut, then - Network accessline 1068
[Keyboard shortcuts](https://help.superhuman.com/hc/en-us/articles/46005701270541-Keyboard-Shortcuts-in-Superhuman-Mail) - Network accessline 1219
- [Keyboard shortcuts](https://help.superhuman.com/hc/en-us/articles/46005701270541-Keyboard-Shortcuts-in-Superhuman-Mail) - Network accessline 1220
- [Autocomplete](https://help.superhuman.com/hc/en-us/articles/46005685782669-Autocomplete) - Network accessline 1221
- [Search](https://help.superhuman.com/hc/en-us/articles/46005672652301-Search) - Network accessline 1222
- [Search in Seconds](https://help.superhuman.com/hc/en-us/articles/46005814266253-Search-in-Seconds) - Network accessline 1223
- [Offline Access](https://help.superhuman.com/hc/en-us/articles/46005499629325-Offline-Access) - Network accessline 1224
- [Undo](https://help.superhuman.com/hc/en-us/articles/46005666743309-Undo) - Network accessline 1225
- [Mark Done](https://help.superhuman.com/hc/en-us/articles/47439134613773-Mark-Done) - Network accessline 1226
- [Attachments](https://help.superhuman.com/hc/en-us/articles/46005568142989-Attachments) - Network accessline 1227
- [Labels](https://help.superhuman.com/hc/en-us/articles/46005736546061-Labels) - Network accessline 1228
- [Folders](https://help.superhuman.com/hc/en-us/articles/46005732666253-Folders) - Network accessline 1229
- [Aliases](https://help.superhuman.com/hc/en-us/articles/46005743269901-Alias) - Network accessline 1230
- [Signatures](https://help.superhuman.com/hc/en-us/articles/46005771841933-Signatures) - Network accessline 1231
- [International keyboard shortcuts](https://help.superhuman.com/hc/en-us/articles/46005584339597-Shortcuts-for-International-Keyboards) - Network accessline 1232
- [Managing Accounts](https://help.superhuman.com/hc/en-us/articles/46005777934733-Managing-Accounts) - Network accessline 1233
- [Customizing Swipes and Triage Bar](https://help.superhuman.com/hc/en-us/articles/46005742942861-Customizing-Swipes-and-Triage-Bar) - Network accessline 1234
- [Remind Me](https://help.superhuman.com/hc/en-us/articles/46005666142733-Remind-Me) - Network accessline 1235
- [Reminders on Autopilot](https://help.superhuman.com/hc/en-us/articles/46005807905421-Reminders-on-Autopilot) - Network accessline 1236
- [Smart Send](https://help.superhuman.com/hc/en-us/articles/46005572688525-Smart-Send) - Network accessline 1237
- [Auto Bcc](https://help.superhuman.com/hc/en-us/articles/46005654497549-Auto-Bcc) - Network accessline 1238
- [Mobile navigation](https://help.superhuman.com/hc/en-us/articles/46005719737357-Mobile-Navigation) - Network accessline 1239
- [Failed sends](https://help.superhuman.com/hc/en-us/articles/46005543693581-Failed-Sends) - Network accessline 1240
- [Quick Quote](https://help.superhuman.com/hc/en-us/articles/46005692763661-Quick-Quote) - Network accessline 1241
- [Instant Reply](https://help.superhuman.com/hc/en-us/articles/46005583725709-Instant-Reply) - Network accessline 1242
- [Shared Conversations and Team Comments](https://help.superhuman.com/hc/en-us/articles/46005593675917-Shared-Conversations-and-Team-Comments) - Network accessline 1243
- [Custom Split Inbox](https://help.superhuman.com/hc/en-us/articles/46005636204941-Custom-Split-Inbox) - Network accessline 1244
- [Your AI Assistant](https://help.superhuman.com/hc/en-us/articles/46005792429965-Your-AI-Assistant) - Network accessline 1245
- [Create Event](https://help.superhuman.com/hc/en-us/articles/46005621734669-Create-Event) - Network accessline 1246
- [Auto Reminders & Auto Drafts](https://help.superhuman.com/hc/en-us/articles/46005658551053-Auto-Reminders-Auto-Drafts) - Network accessline 1247
- [Shared Drafts](https://help.superhuman.com/hc/en-us/articles/46005578703885-Shared-Drafts) - Network accessline 1248
- [Team Features Overview](https://help.superhuman.com/hc/en-us/articles/46005696084109-Team-Features-Overview) - Network accessline 1249
- [Dates, Deadlines, Done](https://help.superhuman.com/hc/en-us/articles/46005854169357-Dates-Deadlines-Done) - Network accessline 1250
- [Read Statuses and Recent Opens Feed](https://help.superhuman.com/hc/en-us/articles/46005603745293-Read-Statuses-and-Recent-Opens-Feed) - Network accessline 1251
- [Email Notifications](https://help.superhuman.com/hc/en-us/articles/46005802618765-Email-Notifications) - Network accessline 1252
- [Reply to Email on Mobile](https://help.superhuman.com/hc/en-us/articles/46005712692877-Reply-to-Email-on-Mobile) - Network accessline 1253
- [Contact Pane](https://help.superhuman.com/hc/en-us/articles/46005778939789-Contact-Pane) - Network accessline 1254
- [Snippets](https://help.superhuman.com/hc/en-us/articles/46005686571149-Snippets) - Network accessline 1255
- [Dealing with Unwanted Emails](https://help.superhuman.com/hc/en-us/articles/46005635358349-Dealing-with-Unwanted-Emails)
templates/mail/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/plan/.agents/skills/actions/SKILL.md
templates/plan/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/plan/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/plan/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/plan/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/plan/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/plan/.agents/skills/visual-plan/SKILL.md
- Network accessline 509
separate slash commands. Pass `--no-connect` to register the connector without authenticating, then run `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 510
authenticating, then run `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` whenever you are ready, or choose a narrower `--client`. Auth and MCP tool - Network accessline 511
`npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 535
Hosted default: connect `https://plan.agent-native.com/mcp`. Do
templates/plan/.agents/skills/visual-plan/references/connection.md
templates/plan/.agents/skills/visual-plan/references/local-files.md
- Network accessline 76
is `false`, verify used the offline lint because the app URL was remote or the local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on - Network accessline 77
local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the - Network accessline 78
`--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the concrete validation error.
templates/plan/.agents/skills/visual-recap/SKILL.md
- Network accessline 264
A recap lives only in the database of the MCP that created it. A separately running local dev server (e.g. `http://localhost:8081`) has its OWN database and - Network accessline 265
running local dev server (e.g. `http://localhost:8081`) has its OWN database and - Network accessline 273
`visualUrl` in the returned `plan.mdx` frontmatter, else `url`/`path` resolved against the MCP server's own origin (for the hosted MCP that is `https://plan.agent-native.com`). This always points at the database that has - Network accessline 274
resolved against the MCP server's own origin (for the hosted MCP that is `https://plan.agent-native.com`). This always points at the database that has the plan. - Network accessline 275
`https://plan.agent-native.com`). This always points at the database that has - Network accessline 277
2. Use a `localhost`/dev origin ONLY when the recap was created through a Plan MCP bound to that same origin — i.e. that MCP's url is `http://localhost:<port>/mcp`. Creating through the hosted MCP - Network accessline 278
MCP bound to that same origin — i.e. that MCP's url is `http://localhost:<port>/mcp`. Creating through the hosted MCP and linking to localhost is the exact mismatch that 404s. - Network accessline 279
`http://localhost:<port>/mcp`. Creating through the hosted MCP - Network accessline 280
and linking to localhost is the exact mismatch that 404s. 3. If only a plan id is available, build the MCP origin's absolute URL (hosted: `https://plan.agent-native.com/plans/<id>`) and say it was inferred. - Network accessline 282
(hosted: `https://plan.agent-native.com/plans/<id>`) and say it was inferred. - Network accessline 285
hosted MCP, say so plainly: the local dev server cannot see it. To view a recap on localhost (e.g. to exercise un-deployed local renderer changes), they must connect a LOCAL Plan MCP (`http://localhost:<port>/mcp`) and - Network accessline 286
on localhost (e.g. to exercise un-deployed local renderer changes), they must connect a LOCAL Plan MCP (`http://localhost:<port>/mcp`) and re-create the recap through it so it lands in the local database; offer to do - Network accessline 287
connect a LOCAL Plan MCP (`http://localhost:<port>/mcp`) and
templates/plan/.agents/skills/visual-recap/references/connection.md
templates/plan/.agents/skills/visual-recap/references/local-files.md
- Network accessline 76
is `false`, verify used the offline lint because the app URL was remote or the local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on - Network accessline 77
local Plan app was unavailable. Run a local Plan app and pass `--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the - Network accessline 78
`--app-url http://localhost:8096` for the authoritative check. If the browser hangs on "Loading plan", fetch the `bridgeUrl` from the verify/serve JSON to read the concrete validation error.
templates/slides-2/.agents/skills/actions/SKILL.md
templates/slides-2/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/slides-2/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/slides-2/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/slides-2/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/slides-2/.agents/skills/image-generation-via-a2a/SKILL.md
- Network accessline 97

templates/slides-2/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/slides/.agents/skills/actions/SKILL.md
templates/slides/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/slides/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/slides/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/slides/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/slides/.agents/skills/image-generation-via-a2a/SKILL.md
- Network accessline 101

templates/slides/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/slides/.agents/skills/slide-editing/references/video.md
templates/tasks/.agents/skills/actions/SKILL.md
templates/tasks/.agents/skills/actions/references/action-fields.md
- Network accessline 38
_agentImages: [ // Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, - Network accessline 39
// Either a public https URL (preferred — the provider fetches it)… { url: "https://cdn.example.com/previews/dash-1.png", label: "overview" }, // …or base64 without a data: prefix (mediaType required; a full - Network accessline 40
{ url: "https://cdn.example.com/previews/dash-1.png", label: "overview" },
templates/tasks/.agents/skills/adding-a-feature/SKILL.md
- Network accessline 194
- **Actions without AGENTS.md** — The actions exist but the agent doesn't know about them because they're not documented. The agent reinvents solutions instead of using the actions. - **Duplicate API routes** — Creating `/api/` routes for o - Network accessline 195
- **Duplicate API routes** — Creating `/api/` routes for operations that actions already handle, including pass-through routes that just call or repackage an action. Use `useActionQuery`/`useActionMutation` instead. - **Raw client route cal - Network accessline 196
- **Raw client route calls** — Teaching or adding `fetch("/_agent-native/...")`, - Network accessline 197
`fetch(agentNativePath(...))`, or template `/api/*` calls in components for
templates/tasks/.agents/skills/efficient-fable/assets/fable-orchestrator.excalidraw
templates/tasks/.agents/skills/external-agents/SKILL.md
- Network accessline 25
An agent-native app is reachable by any MCP-compatible host. Keep setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 26
setup simple: for workspace or cross-app access, add one remote MCP connector: `https://dispatch.agent-native.com/mcp`. Dispatch's Agents page controls whether that single connector reaches all apps or only selected apps, - Network accessline 27
`https://dispatch.agent-native.com/mcp`. Dispatch's Agents page - Network accessline 148
https://dispatch.agent-native.com/mcp - Network accessline 159
https://mail.agent-native.com/mcp - Network accessline 177
npx @agent-native/core@latest connect https://dispatch.agent-native.com - Network accessline 179
npx @agent-native/core@latest connect https://mail.agent-native.com - Network accessline 197
npx -y @agent-native/core@latest reconnect https://dispatch.agent-native.com --client codex - Network accessline 199
npx -y @agent-native/core@latest connect reconnect https://dispatch.agent-native.com --client codex - Access to secretsline 452
[Credential-related source omitted] - Network accessline 481
- Do connect local/fallback clients to Dispatch with `npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Network accessline 482
`npx @agent-native/core@latest connect https://dispatch.agent-native.com`; - Access to secretsline 513
[Credential-related source omitted]
templates/tasks/.agents/skills/external-agents/references/mcp-apps-embedding.md
- Network accessline 168
Microsoft 365 Copilot; Cursor (2.6+); Goose; Postman; MCPJam; Codex Desktop (26.616+); and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 169
and T3 Code (in Codex threads). Desktop Code rendering is observed in [widget bug reports](https://github.com/anthropics/claude-code/issues/103809). Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. - Network accessline 170
[widget bug reports](https://github.com/anthropics/claude-code/issues/103809). - Network accessline 171
Treat the Claude Code terminal, Cowork, and Code on claude.ai web as link-out hosts. [Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), and using it on the same connector can make inline apps dis - Network accessline 172
[Cowork doesn't declare MCP Apps yet](https://github.com/anthropics/claude-code/issues/103924), - Network accessline 174
you reconnect the connector. Host support varies by plan, release channel, and client version, so keep the deep link fallback. See the community-maintained [MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix). - Network accessline 176
[MCP client matrix](https://modelcontextprotocol.io/extensions/client-matrix).
templates/tasks/.agents/skills/secrets/SKILL.md
- Access to secretsline 46
[Credential-related source omitted] - Network accessline 47
label: "Example service API key", description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", - Network accessline 48
description: "Connects this app to the example service.", docsUrl: "https://example.com/api-keys", scope: "user", - Network accessline 49
docsUrl: "https://example.com/api-keys",
templates/tasks/.agents/skills/visual-plan/SKILL.md
- Network accessline 150
registers zero tools), not auth. Stop and give the user the exact restore step for their current client: in Codex/Codex Desktop run `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 151
for their current client: in Codex/Codex Desktop run `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` and start a new Codex session; in Claude Code run `/mcp` and choose - Network accessline 152
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 517
`/visual-plan`, not separate slash commands. Pass `--no-connect` to register the connector without authenticating, then run `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 518
the connector without authenticating, then run `npx @agent-native/core@latest connect https://plan.agent-native.com --client all` whenever you are ready, or choose a narrower `--client`. Auth and MCP tool - Network accessline 519
`npx @agent-native/core@latest connect https://plan.agent-native.com --client all` - Network accessline 535
do not keep retrying the tool. Stop and give the user the reconnect step for the client they are using: Codex/Codex Desktop should run `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 536
client they are using: Codex/Codex Desktop should run `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` and start a new Codex session; Claude Code should run `/mcp` and choose - Network accessline 537
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 546
Hosted default: connect `https://plan.agent-native.com/_agent-native/mcp`. Do
templates/tasks/.agents/skills/visual-recap/SKILL.md
- Network accessline 134
authenticate. Stop and tell the user how to restore it for their current client: in Codex/Codex Desktop, run `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 135
Codex/Codex Desktop, run `npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` and start a new Codex session; in Claude Code, run `/mcp` and choose - Network accessline 136
`npx -y @agent-native/core@latest reconnect https://plan.agent-native.com --client codex` - Network accessline 354
A recap lives only in the database of the MCP that created it. A separately running local dev server (e.g. `http://localhost:8081`) has its OWN database and - Network accessline 355
running local dev server (e.g. `http://localhost:8081`) has its OWN database and - Network accessline 363
`visualUrl` in the returned `plan.mdx` frontmatter, else `url`/`path` resolved against the MCP server's own origin (for the hosted MCP that is `https://plan.agent-native.com`). This always points at the database that has - Network accessline 364
resolved against the MCP server's own origin (for the hosted MCP that is `https://plan.agent-native.com`). This always points at the database that has the plan. - Network accessline 365
`https://plan.agent-native.com`). This always points at the database that has - Network accessline 367
2. Use a `localhost`/dev origin ONLY when the recap was created through a Plan MCP bound to that same origin — i.e. that MCP's url is `http://localhost:<port>/_agent-native/mcp`. Creating through the hosted MCP - Network accessline 368
MCP bound to that same origin — i.e. that MCP's url is `http://localhost:<port>/_agent-native/mcp`. Creating through the hosted MCP and linking to localhost is the exact mismatch that 404s. - Network accessline 369
`http://localhost:<port>/_agent-native/mcp`. Creating through the hosted MCP - Network accessline 370
and linking to localhost is the exact mismatch that 404s. 3. If only a plan id is available, build the MCP origin's absolute URL (hosted: `https://plan.agent-native.com/plans/<id>`) and say it was inferred. - Network accessline 372
(hosted: `https://plan.agent-native.com/plans/<id>`) and say it was inferred. - Network accessline 375
hosted MCP, say so plainly: the local dev server cannot see it. To view a recap on localhost (e.g. to exercise un-deployed local renderer changes), they must connect a LOCAL Plan MCP (`http://localhost:<port>/_agent-native/mcp`) and - Network accessline 376
on localhost (e.g. to exercise un-deployed local renderer changes), they must connect a LOCAL Plan MCP (`http://localhost:<port>/_agent-native/mcp`) and re-create the recap through it so it lands in the app's PostgreSQL database; offer t - Network accessline 377
connect a LOCAL Plan MCP (`http://localhost:<port>/_agent-native/mcp`) and
How the signals fit together
Facts from a static read of the files at the pinned commit. They describe what the code does, not why.
Some files could not be read; the facts below still hold.
Signal confidence: 500 facts
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/design-exploration/SKILL.md:95Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/design-exploration/SKILL.md:108Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/design-exploration/SKILL.md:109Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/design-exploration/SKILL.md:110Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/design-exploration/SKILL.md:111Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/design-exploration/SKILL.md:112Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/design-exploration/SKILL.md:113Low confidence
- Network access in prose of documentation.agents/plugins/agent-native-design/skills/visual-edit/README.md:55Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:18Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:21Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:57Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:73Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:74Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:75Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:78Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:79Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:80Low confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:99High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:100High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:101High confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:175Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:176Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:177Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:198Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:200Low confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:279High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:280High confidence
- Access to secrets in code of the SKILL.md: a variable, placeholder or call, not a stored value.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:280Low confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:291High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:303High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:315High confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:318Low confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:319Low confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:326High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:371High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:372High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:373High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:374High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:394High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:395High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:396High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:397High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:431High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:432High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:433High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:447High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:448High confidence
- Network access in code of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:449High confidence
- Network access in prose of the SKILL.md.agents/plugins/agent-native-design/skills/visual-edit/SKILL.md:470Low confidence
- Network access in code of documentation.agents/plugins/agent-native-turn-into-app/skills/turn-into-app/references/attachments.md:39Low confidence
and 450 more
Only the first 500 facts are listed.
Works with
| Agent | Status | Evidence | Install to |
|---|---|---|---|
| Claude Code | Native | Manifest .agents/plugins/agent-native-design/.claude-plugin/plugin.json | .claude/skills/<skill-name>/ |
| Codex | Native | Manifest AGENTS.md | .agents/skills/<skill-name>/ |
| Cursor | Compatible | Standard SKILL.md community-templates/account-expert/.agents/skills/actions/SKILL.md | .cursor/skills/<skill-name>/ |
| Gemini CLI | Native | Path .gemini/settings.json | .gemini/skills/<skill-name>/ |
| GitHub Copilot | Compatible | Standard SKILL.md community-templates/account-expert/.agents/skills/actions/SKILL.md | .github/skills/<skill-name>/ |
| OpenCode | Compatible | Standard SKILL.md community-templates/account-expert/.agents/skills/actions/SKILL.md | .opencode/skills/<skill-name>/ |
- Native The repository ships a manifest, folder or frontmatter field for this agent.
- Compatible A standard SKILL.md that this agent can read.
- Unknown No evidence either way.
Derived automatically from paths, manifests and frontmatter. Not tested on each agent.
Explore alternatives and comparisons
Add this badge to your README
It updates every night and links to this page.
[](https://hotskills.dev/skill/builderio-agent-native/)