#711 by GitHub stars
langchain-ai/langchain-skills
INVOKE THIS SKILL when building ANY Deep Agents application. Covers create_deep_agent(), harness architecture, SKILL.md format, and configuration options.
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
config/skills/deep-agents-core/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/deep-agents-memory/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/deep-agents-orchestration/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/deepagents-python-quickstart/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/deepagents-typescript-quickstart/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/ecosystem-primer/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/eval-engineering/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langchain-dependencies/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langchain-fundamentals/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langchain-middleware/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langchain-python-quickstart/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langchain-rag/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langchain-typescript-quickstart/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langgraph-cli/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langgraph-decision-models/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langgraph-fundamentals/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langgraph-human-in-the-loop/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langgraph-persistence/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langgraph-python-quickstart/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/langgraph-typescript-quickstart/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/managed-deep-agents/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
config/skills/swarm/SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
126c384d8552 → 126c384d8552
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → 126c384d8552
No earlier version was observed for comparison.
View on GitHubAdd to my skills
Verified commit 126c384d8552dcb476b226c244172186cc3f4328
git clone --filter=blob:none --no-checkout https://github.com/langchain-ai/langchain-skills.git .hotskills/langchain-ai-langchain-skills && git -C .hotskills/langchain-ai-langchain-skills checkout --detach 126c384d8552dcb476b226c244172186cc3f4328
# Copy each selected skill directory from .hotskills/langchain-ai-langchain-skills into .claude/skills/<skill-name>/git clone --filter=blob:none --no-checkout https://github.com/langchain-ai/langchain-skills.git .hotskills/langchain-ai-langchain-skills && git -C .hotskills/langchain-ai-langchain-skills checkout --detach 126c384d8552dcb476b226c244172186cc3f4328
# Copy each selected skill directory from .hotskills/langchain-ai-langchain-skills into .agents/skills/<skill-name>/1.3kGitHub stars
7stars in 7 days
22skills in this repository
4commits in 30 days
Heat
4.1- Star growth, last 7 daysAhead of 74% of tracked repositories
- Star growth, last 24 hoursAhead of 5% of tracked repositories
- Rank movement, last 7 daysAhead of 19% of tracked repositories
- Commits, last 30 daysAhead of 67% 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
- deep-agents-core
config/skills/deep-agents-core - deep-agents-memory
config/skills/deep-agents-memory - deep-agents-orchestration
config/skills/deep-agents-orchestration - deepagents-python-quickstart
config/skills/deepagents-python-quickstart - deepagents-typescript-quickstart
config/skills/deepagents-typescript-quickstart - ecosystem-primer
config/skills/ecosystem-primer - eval-engineering
config/skills/eval-engineering - langchain-dependencies
config/skills/langchain-dependencies - langchain-fundamentals
config/skills/langchain-fundamentals - langchain-middleware
config/skills/langchain-middleware - langchain-python-quickstart
config/skills/langchain-python-quickstart - langchain-rag
config/skills/langchain-rag - langchain-typescript-quickstart
config/skills/langchain-typescript-quickstart - langgraph-cli
config/skills/langgraph-cli - langgraph-decision-models
config/skills/langgraph-decision-models - langgraph-fundamentals
config/skills/langgraph-fundamentals - langgraph-human-in-the-loop
config/skills/langgraph-human-in-the-loop - langgraph-persistence
config/skills/langgraph-persistence - langgraph-python-quickstart
config/skills/langgraph-python-quickstart - langgraph-typescript-quickstart
config/skills/langgraph-typescript-quickstart - managed-deep-agents
config/skills/managed-deep-agents - swarm
config/skills/swarm
How it works
deep-agents-core
When your agent uses it
INVOKE THIS SKILL when building ANY Deep Agents application. Covers create_deep_agent(), harness architecture, SKILL.md format, and configuration options.
What it loads into context
Frontmatter
41 tokensAlways available to the agent
SKILL.md
2,869 tokensWhen the skill is activated
Reference documents
79,537 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- SKILL.md Format
Scripts executed separately
../eval-engineering/references/multi-turn-simulation/harbor_example.py../eval-engineering/references/multi-turn-simulation/model_user.py../eval-engineering/references/multi-turn-simulation/runner.py../eval-engineering/scripts/compare_tool_schemas.py../eval-engineering/scripts/snapshot_sqlite_state.py../swarm/scripts/batching.ts../swarm/scripts/executor.ts../swarm/scripts/filter.ts../swarm/scripts/index.ts../swarm/scripts/interpolate.ts../swarm/scripts/table.ts../swarm/scripts/types.ts../swarm/scripts/utils.ts
Files
../../../.claude-plugin/marketplace.jsonasset488 bytes../../../.claude-plugin/plugin.jsonasset427 bytesSKILL.mdcore12,286 bytesSKILL.mdcore12,286 bytesSKILL.mdcore12,286 bytesSKILL.mdcore12,286 bytesSKILL.mdcore12,286 bytesSKILL.mdcore12,286 bytesSKILL.mdcore12,286 bytesSKILL.mdcore12,286 bytesSKILL.mdcore12,286 bytesSKILL.mdcore12,286 bytes../deep-agents-memory/SKILL.mddoc9,958 bytes../deep-agents-orchestration/SKILL.mddoc13,736 bytes../deepagents-python-quickstart/SKILL.mddoc2,208 bytes../deepagents-typescript-quickstart/SKILL.mddoc2,203 bytes../ecosystem-primer/SKILL.mddoc8,124 bytes../eval-engineering/SKILL.mddoc15,945 bytes../eval-engineering/agents/openai.yamlasset378 bytes../eval-engineering/agents/openai.yamlasset378 bytes../eval-engineering/assets/task/Task.md.templateasset1,872 bytes../eval-engineering/assets/world-skill/SKILL.md.templateasset2,987 bytes../eval-engineering/references/calibration.mddoc3,190 bytes../eval-engineering/references/discovery.mddoc12,347 bytes../eval-engineering/references/environment-building.mddoc4,125 bytes../eval-engineering/references/examples/service-desk.mddoc2,989 bytes../eval-engineering/references/harbor.mddoc6,958 bytes../eval-engineering/references/multi-turn-simulation/guide.mddoc5,177 bytes../eval-engineering/references/multi-turn-simulation/harbor_example.pyruns10,278 bytes../eval-engineering/references/multi-turn-simulation/model_user.pyruns5,117 bytes../eval-engineering/references/multi-turn-simulation/runner.pyruns7,155 bytes../eval-engineering/references/patterns.mddoc37,633 bytes../eval-engineering/references/synthetic-data.mddoc2,953 bytes../eval-engineering/references/task-design.mddoc5,657 bytes../eval-engineering/references/task-implementation.mddoc8,435 bytes../eval-engineering/references/verifier-design.mddoc3,048 bytes../eval-engineering/references/world-knowledge.mddoc11,575 bytes../eval-engineering/scripts/compare_tool_schemas.pyruns14,119 bytes../eval-engineering/scripts/snapshot_sqlite_state.pyruns5,375 bytes../langchain-dependencies/SKILL.mddoc14,594 bytes../langchain-fundamentals/SKILL.mddoc12,524 bytes../langchain-middleware/SKILL.mddoc10,768 bytes../langchain-python-quickstart/SKILL.mddoc1,479 bytes../langchain-rag/SKILL.mddoc15,441 bytes../langchain-typescript-quickstart/SKILL.mddoc1,513 bytes../langgraph-cli/SKILL.mddoc11,190 bytes../langgraph-decision-models/SKILL.mddoc9,398 bytes../langgraph-decision-models/references/conversion-playbook.mddoc6,244 bytes../langgraph-fundamentals/SKILL.mddoc5,903 bytes../langgraph-fundamentals/references/python.mddoc9,178 bytes../langgraph-fundamentals/references/typescript.mddoc9,058 bytes../langgraph-human-in-the-loop/SKILL.mddoc16,411 bytes../langgraph-persistence/SKILL.mddoc18,529 bytes../langgraph-python-quickstart/SKILL.mddoc1,827 bytes../langgraph-typescript-quickstart/SKILL.mddoc1,832 bytes../managed-deep-agents/SKILL.mddoc34,989 bytes../swarm/SKILL.mddoc12,048 bytes../swarm/scripts/batching.tsruns8,897 bytes../swarm/scripts/executor.tsruns4,694 bytes../swarm/scripts/filter.tsruns1,818 bytes../swarm/scripts/index.tsruns10,746 bytes../swarm/scripts/interpolate.tsruns2,319 bytes../swarm/scripts/table.tsruns16,109 bytes../swarm/scripts/types.tsruns7,598 bytes../swarm/scripts/utils.tsruns892 bytes
Read SKILL.md
<overview>
Deep Agents are an opinionated agent framework built on LangChain/LangGraph with built-in middleware:
- **Task Planning**: TodoListMiddleware for breaking down complex tasks
- **Context Management**: Filesystem tools with pluggable backends
- **Task Delegation**: SubAgent middleware for spawning specialized agents
- **Long-term Memory**: Persistent storage across threads via Store
- **Human-in-the-loop**: Approval workflows for sensitive operations
- **Skills**: On-demand loading of specialized capabilities
The agent harness provides these capabilities automatically - you configure, not implement.
</overview>
<when-to-use>deep-agents-memory
When your agent uses it
INVOKE THIS SKILL when your Deep Agent needs memory, persistence, or filesystem access. Covers StateBackend (ephemeral), StoreBackend (persistent), FilesystemMiddleware, and CompositeBackend for routing.
What it loads into context
Frontmatter
51 tokensAlways available to the agent
SKILL.md
2,359 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
Files
SKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytesSKILL.mdcore9,958 bytes
Read SKILL.md
<overview>
Deep Agents use pluggable backends for file operations and memory:
**Short-term (StateBackend)**: Persists within a single thread, lost when thread ends
**Long-term (StoreBackend)**: Persists across threads and sessions
**Hybrid (CompositeBackend)**: Route different paths to different backends
FilesystemMiddleware provides tools: `ls`, `read_file`, `write_file`, `edit_file`, `glob`, `grep`
</overview>
<backend-selection>
| Use Case | Backend | Why |
|----------|---------|-----|deep-agents-orchestration
When your agent uses it
INVOKE THIS SKILL when using subagents, task planning, or human approval in Deep Agents. Covers SubAgentMiddleware, TodoList for planning, and HITL interrupts.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
3,241 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
- Subagents (Task Delegation)
- TodoList (Task Planning)
- Human-in-the-Loop (Approval Workflows)
Files
SKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytesSKILL.mdcore13,736 bytes
Read SKILL.md
<overview>
Deep Agents include three orchestration capabilities:
1. **SubAgentMiddleware**: Delegate work via `task` tool to specialized agents
2. **TodoListMiddleware**: Plan and track tasks via `write_todos` tool
3. **HumanInTheLoopMiddleware**: Require approval before sensitive operations
All three are automatically included in `create_deep_agent()`.
</overview>
---
## Subagents (Task Delegation)
deepagents-python-quickstart
When your agent uses it
Scaffold a minimal local Deep Agent in Python by following the official quickstart, using provider-native web search instead of Tavily. Use when the user wants to quickly build or try a Deep Agent locally.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
544 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
- Local setup constraints
Files
SKILL.mdcore2,208 bytes
Read SKILL.md
# Deep Agents Python quickstart
Follow the live docs — do not invent an alternate API from memory:
**https://docs.langchain.com/oss/python/deepagents/quickstart**
Fetch that page (Docs MCP or HTTP) and implement the research-agent shape it shows (`create_deep_agent`, research system prompt, invoke with a research question like “What is LangGraph?”).
## Local setup constraints
Apply these on top of the quickstart (they keep setup minimal and model-agnostic):
1. **Ask** which provider/model to use. Showcase that Deep Agents are model-agnostic. Suggested prompt:
deepagents-typescript-quickstart
When your agent uses it
Scaffold a minimal local Deep Agent in TypeScript by following the official quickstart, using provider-native web search instead of Tavily. Use when the user wants to quickly build or try a Deep Agent locally.
What it loads into context
Frontmatter
55 tokensAlways available to the agent
SKILL.md
534 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
- Local setup constraints
Files
SKILL.mdcore2,203 bytes
Read SKILL.md
# Deep Agents TypeScript quickstart
Follow the live docs — do not invent an alternate API from memory:
**https://docs.langchain.com/oss/javascript/deepagents/quickstart**
Fetch that page (Docs MCP or HTTP) and implement the research-agent shape it shows (`createDeepAgent`, research system prompt, invoke with a research question like “What is LangGraph?”). Requires Node 22+.
## Local setup constraints
Apply these on top of the quickstart (they keep setup minimal and model-agnostic):
1. **Ask** which provider/model to use. Showcase that Deep Agents are model-agnostic. Suggested prompt:
ecosystem-primer
When your agent uses it
INVOKE FIRST for any LangChain / LangGraph / Deep Agents agent building project before consulting other skills or writing any agent code. Required starting point for up to date info on framework selection (LangChain vs LangGraph vs Deep Agents vs hybrid composition), agent patterns, install, environment setup, and which skill to load next.
What it loads into context
Frontmatter
75 tokensAlways available to the agent
SKILL.md
1,841 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
- Step 1 — Choose Your Tool
- Tool Profiles
- Mixing Layers
- Step 2 — Set Environment Variables
- Step 3 — How the Docs Work
- Step 4 — Load the Right Skill Next
Files
SKILL.mdcore8,124 bytesSKILL.mdcore8,124 bytes
Read SKILL.md
<overview>
LangChain Inc. maintains three layered open-source tools for building agents, plus LangSmith for observability. The stack, top-down:
- **Deep Agents** (top layer, *harness*) — batteries-included toolkit built on LangChain + LangGraph. Ships with planning, file management, subagent spawning, and memory out of the box.
- **LangGraph** (middle layer, *runtime*) — low-level orchestration for durable execution, custom control flow, and stateful workflows. LangChain agents run on top of LangGraph.
- **LangChain** (bottom layer, *framework*) — abstractions for models, tools, and the agent loop. Provider-agnostic, easiest to start with.
- **LangSmith** (cross-cutting) — observability and evaluation platform. Framework-agnostic; always recommended alongside any of the above.
Higher layers depend on lower ones, but you don't need to use lower layers directly. Deep Agents gives you LangGraph's durable execution without writing graph code. LangChain gives you models and tools without managing graph edges.
</overview>
---
## Step 1 — Choose Your Tooleval-engineering
When your agent uses it
Inspect an agent repository and optional traces, interview the user, write reviewed Task Specs, build and audit Harbor tasks, and bootstrap reusable project World Knowledge Skills. Use for agent evals, benchmark design, Task generation, controlled Environments, synthetic data, Verifiers, Harbor runs, calibration, or continuous benchmark maintenance.
What it loads into context
Frontmatter
71 tokensAlways available to the agent
SKILL.md
3,342 tokensWhen the skill is activated
Reference documents
21,212 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Flow
- Terms
- Reference routing
- 1. Inspect inputs and existing World knowledge
- 2. Propose and select a Task
- 3. Write and review the Task Spec and World Skill
- 4. Apply Spec2Task
- 5. Run and audit
- 6. Reconcile project World knowledge
- 7. Repeat
- Dependencies, access, and safety
- Complete only when
Scripts executed separately
references/multi-turn-simulation/harbor_example.pyreferences/multi-turn-simulation/model_user.pyreferences/multi-turn-simulation/runner.pyscripts/compare_tool_schemas.pyscripts/snapshot_sqlite_state.py
Files
SKILL.mdcore15,945 bytesSKILL.mdcore15,945 bytesSKILL.mdcore15,945 bytesSKILL.mdcore15,945 bytesSKILL.mdcore15,945 bytesSKILL.mdcore15,945 bytesSKILL.mdcore15,945 bytesSKILL.mdcore15,945 bytesagents/openai.yamlasset378 bytesassets/task/Task.md.templateasset1,872 bytesassets/world-skill/SKILL.md.templateasset2,987 bytesreferences/calibration.mddoc3,190 bytesreferences/discovery.mddoc12,347 bytesreferences/environment-building.mddoc4,125 bytesreferences/examples/service-desk.mddoc2,989 bytesreferences/harbor.mddoc6,958 bytesreferences/multi-turn-simulation/guide.mddoc5,177 bytesreferences/multi-turn-simulation/harbor_example.pyruns10,278 bytesreferences/multi-turn-simulation/model_user.pyruns5,117 bytesreferences/multi-turn-simulation/runner.pyruns7,155 bytesreferences/patterns.mddoc37,633 bytesreferences/synthetic-data.mddoc2,953 bytesreferences/task-design.mddoc5,657 bytesreferences/task-implementation.mddoc8,435 bytesreferences/verifier-design.mddoc3,048 bytesreferences/world-knowledge.mddoc11,575 bytesscripts/compare_tool_schemas.pyruns14,119 bytesscripts/snapshot_sqlite_state.pyruns5,375 bytes
Read SKILL.md
# Eval Engineering
## Flow
- Inspect all inputs first: the repository, Harness, optional traces, existing
Tasks and runs, existing World knowledge, and the human goal. Identify the
source files and skill references that apply before proposing work.
- Create or update the small project World Knowledge Skill from reusable facts
in those inputs. Use it to propose one grounded Task.
- Draft the Task Spec and World Skill together. Show both to the user, keep exact
Task truth only in `Task.md`, and refine both until the user approves them.
- Implement the approved Task, validate its Environment and Verifier, run the
real Harness, inspect the full evidence, and fix only non-agent failures.
- Reconcile the World Skill with what the run proved, then repeat this flow forlangchain-dependencies
When your agent uses it
INVOKE THIS SKILL when setting up a new project or when asked about package versions, installation, or dependency management for LangChain, LangGraph, LangSmith, or Deep Agents. Covers required packages, minimum versions, environment requirements, versioning best practices, and common community tool packages for both Python and TypeScript.
What it loads into context
Frontmatter
75 tokensAlways available to the agent
SKILL.md
3,791 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
- Environment Requirements
- Framework Choice
- Core Packages
- Minimal Project Templates
- Versioning Policy & Upgrade Strategy
- Environment Variables
- Common Mistakes
Files
SKILL.mdcore14,594 bytesSKILL.mdcore14,594 bytesSKILL.mdcore14,594 bytesSKILL.mdcore14,594 bytesSKILL.mdcore14,594 bytes
Read SKILL.md
<overview>
The LangChain ecosystem is split into focused, independently-versioned packages. Understanding which packages you need — and their version constraints — prevents incompatibilities and keeps upgrades predictable.
**Key principles:**
- **LangChain 1.0 is the current LTS release.** Always start new projects on 1.0+. LangChain 0.3 is legacy maintenance-only — do not use it for new work.
- **langchain-core** is the shared foundation: always install it explicitly alongside any other package.
- **langchain-community** (Python only) does NOT follow semantic versioning; pin it conservatively.
- **LangGraph vs Deep Agents:** choose one orchestration approach based on your use case — they are alternatives, not a required stack (see [Framework Choice](#framework-choice) below).
- Provider integrations (model, vector store, tools) are installed separately so you only pull in what you use.
</overview>
---
## Environment Requirementslangchain-fundamentals
When your agent uses it
Create LangChain agents with create_agent, define tools, and use middleware for human-in-the-loop and error handling.
What it loads into context
Frontmatter
34 tokensAlways available to the agent
SKILL.md
3,107 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
- Creating Agents with create_agent
- Defining Tools
- Middleware for Agent Control
- Structured Output
- Model Configuration
Files
SKILL.mdcore12,524 bytesSKILL.mdcore12,524 bytesSKILL.mdcore12,524 bytesSKILL.mdcore12,524 bytesSKILL.mdcore12,524 bytesSKILL.mdcore12,524 bytesSKILL.mdcore12,524 bytesSKILL.mdcore12,524 bytesSKILL.mdcore12,524 bytesSKILL.mdcore12,524 bytesSKILL.mdcore12,524 bytes
Read SKILL.md
<oneliner>
Build production agents using `create_agent()`, middleware patterns, and the `@tool` decorator / `tool()` function. When creating LangChain agents, you MUST use create_agent(), with middleware for custom flows. All other alternatives are outdated.
</oneliner>
<create_agent>
## Creating Agents with create_agent
`create_agent()` is the recommended way to build agents. It handles the agent loop, tool execution, and state management.
### Agent Configuration Options
| Parameter | Purpose | Example |
|-----------|---------|---------|
| `model` | LLM to use | `"anthropic:claude-sonnet-4-5"` or model instance |langchain-middleware
When your agent uses it
INVOKE THIS SKILL when you need human-in-the-loop approval, custom middleware, or structured output. Covers HumanInTheLoopMiddleware for human approval of dangerous tool calls, creating custom middleware with hooks, Command resume patterns, and structured output with Pydantic/Zod.
What it loads into context
Frontmatter
66 tokensAlways available to the agent
SKILL.md
2,583 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
- Human-in-the-Loop
- Custom Middleware Hooks
Files
SKILL.mdcore10,768 bytesSKILL.mdcore10,768 bytesSKILL.mdcore10,768 bytesSKILL.mdcore10,768 bytesSKILL.mdcore10,768 bytesSKILL.mdcore10,768 bytesSKILL.mdcore10,768 bytesSKILL.mdcore10,768 bytesSKILL.mdcore10,768 bytesSKILL.mdcore10,768 bytesSKILL.mdcore10,768 bytes
Read SKILL.md
<overview>
Middleware patterns for production LangChain agents:
- **HumanInTheLoopMiddleware** / **humanInTheLoopMiddleware**: Pause before dangerous tool calls for human approval
- **Custom middleware**: Intercept tool calls for error handling, logging, retry logic
- **Command resume**: Continue execution after human decisions (approve, edit, reject)
**Requirements:** Checkpointer + thread_id config for all HITL workflows.
</overview>
---
## Human-in-the-Loop
langchain-python-quickstart
When your agent uses it
Scaffold a minimal local LangChain agent in Python by following the official quickstart. Use when the user wants to quickly build or try a LangChain agent locally.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
365 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
- Local setup constraints
Files
SKILL.mdcore1,479 bytes
Read SKILL.md
# LangChain Python quickstart
Follow the live docs — do not invent an alternate API from memory:
**https://docs.langchain.com/oss/python/langchain/quickstart**
Fetch that page (Docs MCP or HTTP) and implement what it shows (weather agent + `create_agent`).
## Local setup constraints
Apply these on top of the quickstart (they keep setup minimal and model-agnostic):
1. **Ask** which provider/model to use. Showcase that LangChain is model-agnostic. Suggested prompt:
langchain-rag
When your agent uses it
INVOKE THIS SKILL when building ANY retrieval-augmented generation (RAG) system. Covers document loaders, RecursiveCharacterTextSplitter, embeddings (OpenAI), and vector stores (Chroma, FAISS, Pinecone).
What it loads into context
Frontmatter
57 tokensAlways available to the agent
SKILL.md
3,729 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
- Complete RAG Pipeline
- Document Loaders
- Text Splitting
- Vector Stores
- Retrieval
Files
SKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytesSKILL.mdcore15,441 bytes
Read SKILL.md
<overview>
Retrieval Augmented Generation (RAG) enhances LLM responses by fetching relevant context from external knowledge sources.
**Pipeline:**
1. **Index**: Load → Split → Embed → Store
2. **Retrieve**: Query → Embed → Search → Return docs
3. **Generate**: Docs + Query → LLM → Response
**Key Components:**
- **Document Loaders**: Ingest data from files, web, databases
- **Text Splitters**: Break documents into chunks
- **Embeddings**: Convert text to vectors
- **Vector Stores**: Store and search embeddings
</overview>langchain-typescript-quickstart
When your agent uses it
Scaffold a minimal local LangChain agent in TypeScript by following the official quickstart. Use when the user wants to quickly build or try a LangChain agent locally.
What it loads into context
Frontmatter
47 tokensAlways available to the agent
SKILL.md
373 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
- Local setup constraints
Files
SKILL.mdcore1,513 bytes
Read SKILL.md
# LangChain TypeScript quickstart
Follow the live docs — do not invent an alternate API from memory:
**https://docs.langchain.com/oss/javascript/langchain/quickstart**
Fetch that page (Docs MCP or HTTP) and implement what it shows (weather agent + `createAgent`). Requires Node 22+.
## Local setup constraints
Apply these on top of the quickstart (they keep setup minimal and model-agnostic):
1. **Ask** which provider/model to use. Showcase that LangChain is model-agnostic. Suggested prompt:
langgraph-cli
When your agent uses it
INVOKE THIS SKILL when using the langgraph CLI to scaffold, develop, build, or deploy LangGraph applications. Covers langgraph new, dev, build, up, deploy, and langgraph.json configuration.
What it loads into context
Frontmatter
53 tokensAlways available to the agent
SKILL.md
2,674 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
- When to use
- Installation
- Commands
- `langgraph.json` reference
- Typical workflow
- `langgraph dev` vs `langgraph up`
- Gotchas
Files
SKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytesSKILL.mdcore11,190 bytes
Read SKILL.md
<overview>
The `langgraph` CLI manages the full lifecycle of LangGraph applications — from scaffolding a new project to deploying it to LangGraph Platform (LangSmith Deployments).
Key commands:
- **`langgraph new`** — Scaffold a project from a template
- **`langgraph dev`** — Run locally with hot reload (no Docker)
- **`langgraph build`** — Build a Docker image
- **`langgraph up`** — Launch locally via Docker Compose
- **`langgraph deploy`** — Ship to LangGraph Platform
- **`langgraph dockerfile`** — Generate a Dockerfile
All commands (except `new`) read from a `langgraph.json` config file in the project root.
</overview>
langgraph-decision-models
When your agent uses it
INVOKE THIS SKILL when routing a LangGraph agent with a decision model (TypeSafe Jev, SemIf) instead of an LLM, or when auditing an existing agent for LLM calls that only produce a routing decision. Covers langchain-typesafe Noul/Choice/Score, reading answers correctly, threshold design, and LangSmith Gateway wiring.
What it loads into context
Frontmatter
85 tokensAlways available to the agent
SKILL.md
2,261 tokensWhen the skill is activated
Reference documents
1,338 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- Install and wire
- The conversion pattern
- Three traps when reading answers
- Question wording dominates accuracy
- Auditing an existing agent
- Expectations
- Batching with Send
- Related skills
Files
SKILL.mdcore9,398 bytesSKILL.mdcore9,398 bytesSKILL.mdcore9,398 bytesSKILL.mdcore9,398 bytesSKILL.mdcore9,398 bytesSKILL.mdcore9,398 bytesSKILL.mdcore9,398 bytesSKILL.mdcore9,398 bytesSKILL.mdcore9,398 bytesSKILL.mdcore9,398 bytesreferences/conversion-playbook.mddoc6,244 bytes
Read SKILL.md
<overview>
A **decision model** answers typed questions about state and returns probabilities instead of prose. It replaces the common pattern of prompting an LLM, parsing its text, and branching on the result.
- **`Noul(instructions=...)`** — binary question, returns a probability of yes
- **`Choice(instructions=..., criteria={...})`** — picks one label, returns the full distribution plus confidence
- **`Score(instructions=..., criteria=[...])`** — grades against an ordered rubric, returns an expected value plus confidence
`TypeSafeClassifier` is a LangChain `Runnable[ClassifierRequest, ClassifierResponse]`, so it drops into a node like any other runnable. Up to 32 questions share one request and are answered independently — your code combines them.
**Reach for one when** a node generates text only so you can parse a decision out of it: routing, triage, filtering, guardrails, or per-item classification over a batch.
**Do not** reach for one when the node's output is the product (summaries, drafts, code) or when the judgment needs multi-step reasoning. A decision model classifies; it does not think.
</overview>
langgraph-fundamentals
When your agent uses it
INVOKE THIS SKILL when writing ANY LangGraph code. Covers StateGraph, state schemas, nodes, edges, Command, Send, invoke, streaming, and error handling.
What it loads into context
Frontmatter
48 tokensAlways available to the agent
SKILL.md
1,340 tokensWhen the skill is activated
Reference documents
4,606 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- State Management
- Nodes
- Edges
- Command
- Send API
- Running Graphs: Invoke and Stream
- Error Handling
- Core boundaries
- Implementation references
Files
SKILL.mdcore5,903 bytesreferences/python.mddoc9,178 bytesreferences/typescript.mddoc9,058 bytes
Read SKILL.md
<overview>
LangGraph models agent workflows as **directed graphs**:
- **StateGraph**: Main class for building stateful graphs
- **Nodes**: Functions that perform work and update state
- **Edges**: Define execution order (static or conditional)
- **START/END**: Special nodes marking entry and exit points
- **State with Reducers**: Control how state updates are merged
Graphs must be `compile()`d before execution.
</overview>
<design-methodology>
langgraph-human-in-the-loop
When your agent uses it
INVOKE THIS SKILL when implementing human-in-the-loop patterns, pausing for approval, or handling errors in LangGraph. Covers interrupt(), Command(resume=...), approval/validation workflows, and the 4-tier error handling strategy.
What it loads into context
Frontmatter
60 tokensAlways available to the agent
SKILL.md
3,896 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
- Requirements
- Basic Interrupt + Resume
- Approval Workflow
- Validation Loop
- Multiple Interrupts
- Side Effects Before Interrupt Must Be Idempotent
- Command(resume) Warning
- Fixes
Files
SKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytesSKILL.mdcore16,411 bytes
Read SKILL.md
<overview>
LangGraph's human-in-the-loop patterns let you pause graph execution, surface data to users, and resume with their input:
- **`interrupt(value)`** — pauses execution, surfaces a value to the caller
- **`Command(resume=value)`** — resumes execution, providing the value back to `interrupt()`
- **Checkpointer** — required to save state while paused
- **Thread ID** — required to identify which paused execution to resume
</overview>
---
## Requirements
Three things are required for interrupts to work:langgraph-persistence
When your agent uses it
INVOKE THIS SKILL when your LangGraph needs to persist state, remember conversations, travel through history, or configure subgraph checkpointer scoping. Covers checkpointers, thread_id, time travel, Store, and subgraph persistence modes.
What it loads into context
Frontmatter
60 tokensAlways available to the agent
SKILL.md
4,345 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
- Checkpointer Setup
- Thread Management
- State History & Time Travel
- Subgraph Checkpointer Scoping
- Long-Term Memory (Store)
- Fixes
Files
SKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytesSKILL.mdcore18,529 bytes
Read SKILL.md
<overview>
LangGraph's persistence layer enables durable execution by checkpointing graph state:
- **Checkpointer**: Saves/loads graph state at every super-step
- **Thread ID**: Identifies separate checkpoint sequences (conversations)
- **Store**: Cross-thread memory for user preferences, facts
**Two memory types:**
- **Short-term** (checkpointer): Thread-scoped conversation history
- **Long-term** (store): Cross-thread user preferences, facts
</overview>
<checkpointer-selection>
langgraph-python-quickstart
When your agent uses it
Scaffold a minimal local LangGraph agent in Python by following the official quickstart. Use when the user wants to quickly build or try a LangGraph agent locally.
What it loads into context
Frontmatter
45 tokensAlways available to the agent
SKILL.md
450 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
- Local setup constraints
Files
SKILL.mdcore1,827 bytes
Read SKILL.md
# LangGraph Python quickstart
Follow the live docs — do not invent an alternate API from memory:
**https://docs.langchain.com/oss/python/langgraph/quickstart**
Fetch that page (Docs MCP or HTTP) and implement what it shows (calculator / math agent with the Graph API). Prefer the Graph API path over the Functional API unless the user asks otherwise. Skip IPython graph visualization.
## Local setup constraints
Apply these on top of the quickstart (they keep setup minimal and model-agnostic):
1. **Ask** which provider/model to use. Showcase that LangGraph works with any LangChain chat model. Suggested prompt:
langgraph-typescript-quickstart
When your agent uses it
Scaffold a minimal local LangGraph agent in TypeScript by following the official quickstart. Use when the user wants to quickly build or try a LangGraph agent locally.
What it loads into context
Frontmatter
47 tokensAlways available to the agent
SKILL.md
451 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
- Local setup constraints
Files
SKILL.mdcore1,832 bytes
Read SKILL.md
# LangGraph TypeScript quickstart
Follow the live docs — do not invent an alternate API from memory:
**https://docs.langchain.com/oss/javascript/langgraph/quickstart**
Fetch that page (Docs MCP or HTTP) and implement what it shows (calculator / math agent with the Graph API). Prefer the Graph API path over the Functional API unless the user asks otherwise. Skip graph visualization.
## Local setup constraints
Apply these on top of the quickstart (they keep setup minimal and model-agnostic):
1. **Ask** which provider/model to use. Showcase that LangGraph works with any LangChain chat model. Suggested prompt:
managed-deep-agents
When your agent uses it
INVOKE THIS SKILL when building, testing, or deploying Managed Deep Agents in LangSmith. Walks a user through their first agent end to end — interviewing them about what they want to build, mapping it onto what MDA can actually do, then scaffolding and deploying it. Covers the file-based project layout; define_deep_agent / defineDeepAgent; instructions, skills, memory, identity, tools, MCP servers, connections, middleware, sandboxes, schedules, channels, and evals; the mda CLI; GitHub deployments; and Context Hub.
What it loads into context
Frontmatter
126 tokensAlways available to the agent
SKILL.md
8,086 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
- When to use
- 1. Ask what they want to build
- 2. Check the answer against the limits
- 3. Map the answer onto capabilities
- 4. Confirm the shape before writing files
- 5. Scaffold and wire the smallest thing that runs
- 6. Handle keys without touching their secrets
- 7. Run it locally, then deploy
- What MDA cannot do
- Prerequisites
- Project layout
- Define the agent
- Instructions
- Skills
- Memory
- Identity
- Tools
- MCP servers and connections
- Middleware
- Sandboxes
- Schedules
- Channels
- Evals
- CLI reference
- Deploy and Context Hub
- Human-in-the-loop
- Gotchas
Files
SKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytesSKILL.mdcore34,989 bytes
Read SKILL.md
# Managed Deep Agents
## Overview
Managed Deep Agents (MDA) is a hosted runtime for code-first Deep Agents in LangSmith. You author an agent in Python or TypeScript, test it locally with `mda dev`, and deploy it with `mda deploy` or from a connected GitHub repository in LangSmith. It pairs the open-source Deep Agents harness (see [[deep-agents-core]]) with managed infrastructure: durable runs, sandboxes, Context Hub-backed instructions and skills, memory, traces, and hosted LangGraph deployment.
The core idea is that **an agent is a directory**. A file's location determines its role, and the CLI compiles that directory into a managed LangGraph app.
MDA is in **public beta** and runs on **US LangSmith Cloud only**.
## When to use
Use this skill when the user wants to build a Deep Agent in code and run it on LangSmith without operating their own server, or to add tools, MCP servers, connections, middleware, memory, identity, schedules, channels, skills, sandboxes, or evals to one.
swarm
When your agent uses it
Dispatches many independent items in parallel: create a table, fan out to subagents, aggregate results. One row = one unit of work.
What it loads into context
Frontmatter
88 tokensAlways available to the agent
SKILL.md
3,069 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
- Flow
- Choosing a source
- When to use `subagentType`
- Instruction + context
- Structured output
- Batching
- Aggregation
- Chaining passes
- Action-only tasks
- Filtering
- Technical notes
- API Reference
Scripts executed separately
scripts/batching.tsscripts/executor.tsscripts/filter.tsscripts/index.tsscripts/interpolate.tsscripts/table.tsscripts/types.tsscripts/utils.ts
Files
SKILL.mdcore12,048 bytesSKILL.mdcore12,048 bytesSKILL.mdcore12,048 bytesscripts/batching.tsruns8,897 bytesscripts/executor.tsruns4,694 bytesscripts/filter.tsruns1,818 bytesscripts/index.tsruns10,746 bytesscripts/interpolate.tsruns2,319 bytesscripts/table.tsruns16,109 bytesscripts/types.tsruns7,598 bytesscripts/utils.tsruns892 bytes
Read SKILL.md
# Swarm
Process many independent items in parallel. `create` builds a table handle;
`run` fans work out across rows and merges results back. One row = one unit
of work — swarm handles batching automatically.
## Flow
1. **Create.** Build a table from a source — files, a glob pattern, or
pre-parsed records. One row per item. Returns a handle.
2. **Run.** Dispatch an `instruction` template across rows. Results are merged
back into the table. Returns `{ completed, failed, skipped, failures }`.
3. **Aggregate.** Use `rows()` and plain JS to count, filter, or summarize.
Do not spawn additional subagents for aggregation.Skills included
- deep-agents-core
- deep-agents-memory
- deep-agents-orchestration
- deepagents-python-quickstart
- deepagents-typescript-quickstart
- ecosystem-primer
- eval-engineering
- langchain-dependencies
- langchain-fundamentals
- langchain-middleware
- langchain-python-quickstart
- langchain-rag
- langchain-typescript-quickstart
- langgraph-cli
- langgraph-decision-models
- langgraph-fundamentals
- langgraph-human-in-the-loop
- langgraph-persistence
- langgraph-python-quickstart
- langgraph-typescript-quickstart
- managed-deep-agents
- swarm
Repository facts
- License
- MIT
- Created
- Jan 22, 2026
- Last pushed
- Nov 5, 2026
- Forks
- 108
- Open issues
- 2
- Project activity
- Active
Rank over the last 12 weeks
Now #711 · best #504 · down 125 in 30 days
- #699 on Nov 1
- #702 on Nov 2
- #704 on Nov 3
- #707 on Nov 4
- #710 on Nov 5
- #711 since Nov 6
75 earlier periods
- #513 on Aug 2
- #510 on Aug 7
- #503 on Aug 8
- #505 on Aug 9
- #523 on Aug 10
- #522 on Aug 11
- #520 on Aug 12
- #523 on Aug 13
- #522 on Aug 14
- #504 on Aug 15
- #516 on Aug 16
- #518 on Aug 17
- #536 on Aug 18
- #528 on Aug 20
- #531 on Aug 21
- #516 on Aug 22
- #536 on Aug 23
- #538 on Aug 24
- #547 on Aug 25
- #533 on Aug 26
- #546 on Aug 27
- #547 on Aug 28
- #527 on Aug 29
- #519 on Aug 30
- #546 on Aug 31
- #551 on Sep 1
- #543 on Sep 2
- #546 on Sep 3
- #529 on Sep 4
- #540 on Sep 6
- #561 on Sep 8
- #546 on Sep 9
- #551 on Sep 10
- #543 on Sep 11
- #538 on Sep 12
- #540 on Sep 13
- #562 on Sep 14
- #548 on Sep 15
- #560 from Sep 16 to Sep 18
- #559 on Sep 20
- #567 on Sep 21
- #576 on Sep 22
- #575 on Sep 23
- #556 on Sep 24
- #549 from Sep 25 to Sep 26
- #572 on Sep 27
- #576 on Sep 28
- #597 on Sep 29
- #584 on Sep 30
- #567 on Oct 1
- #576 on Oct 2
- #567 on Oct 4
- #575 on Oct 5
- #586 on Oct 6
- #614 on Oct 8
- #599 on Oct 9
- #586 on Oct 10
- #588 on Oct 11
- #602 from Oct 12 to Oct 13
- #616 on Oct 14
- #611 on Oct 15
- #619 on Oct 16
- #604 on Oct 17
- #599 on Oct 18
- #627 on Oct 19
- #635 on Oct 20
- #642 on Oct 21
- #624 on Oct 22
- #617 on Oct 23
- #605 on Oct 24
- #682 on Oct 25
- #692 on Oct 28
- #693 on Oct 29
- #696 on Oct 30
- #698 on Oct 31
Before you install
Network or broad shellAutomated signals, not a guarantee
Checked Nov 3, 2026
Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.
Show 167 notes
config/skills/deep-agents-core/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-core/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-core/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-core/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-core/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-core/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-core/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-core/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-core/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/eval-engineering/agents/openai.yaml:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-memory/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-memory/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-memory/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-memory/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-memory/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-memory/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-memory/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-memory/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-memory/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzedconfig/skills/deep-agents-memory/SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzed
- No broad tool permissions
- No dangerous shell commands
- 196 network referencesin 15 files: config/skills/eval-engineering/references/patterns.md:25 config/skills/managed-deep-agents/SKILL.md:138 config/skills/langgraph-decision-models/SKILL.md:43 and 12 more
- No hidden or obfuscated text
- 2 instruction override signalsin 1 file: config/skills/managed-deep-agents/SKILL.md:271
- 70 references to secretsin 6 files: config/skills/langchain-dependencies/SKILL.md:303 config/skills/managed-deep-agents/SKILL.md:109 config/skills/langgraph-cli/SKILL.md:118 and 3 more
- No requests to vote in Hot
Show all 294 signals by file
.claude-plugin/marketplace.json
- Network accessline 13
"source": "./", "description": "Agent skills for building agents with LangChain, LangGraph, and Deep Agents", "homepage": "https://github.com/langchain-ai/langchain-skills" - Network accessline 14
"description": "Agent skills for building agents with LangChain, LangGraph, and Deep Agents", "homepage": "https://github.com/langchain-ai/langchain-skills" } - Network accessline 15
"homepage": "https://github.com/langchain-ai/langchain-skills"
.claude-plugin/plugin.json
- Network accessline 5
"author": { "name": "LangChain", "url": "https://langchain.com" - Network accessline 6
"name": "LangChain", "url": "https://langchain.com" }, - Network accessline 7
"url": "https://langchain.com" - Network accessline 8
}, "repository": "https://github.com/langchain-ai/langchain-skills", "license": "MIT", - Network accessline 9
"repository": "https://github.com/langchain-ai/langchain-skills",
config/skills/deepagents-python-quickstart/SKILL.md
config/skills/deepagents-typescript-quickstart/SKILL.md
config/skills/ecosystem-primer/SKILL.md
- Access to secretsline 115
[Credential-related source omitted] - Access to secretsline 115
[Credential-related source omitted] - Access to secretsline 120
[Credential-related source omitted] - Access to secretsline 120
[Credential-related source omitted] - Network accessline 154
**If the MCP server is not available**, use the `llms.txt` index: 1. Fetch `https://docs.langchain.com/llms.txt` — structured list of all pages with descriptions - Network accessline 154
**If the MCP server is not available**, use the `llms.txt` index: 1. Fetch `https://docs.langchain.com/llms.txt` — structured list of all pages with descriptions - Network accessline 155
1. Fetch `https://docs.langchain.com/llms.txt` — structured list of all pages with descriptions - Network accessline 155
1. Fetch `https://docs.langchain.com/llms.txt` — structured list of all pages with descriptions
config/skills/eval-engineering/references/multi-turn-simulation/harbor_example.py
config/skills/eval-engineering/references/multi-turn-simulation/model_user.py
config/skills/eval-engineering/references/multi-turn-simulation/runner.py
config/skills/eval-engineering/references/patterns.md
- Network accessline 25
| Benchmark | Domain | Design worth studying | |---|---|---| | [AutomationBench](https://github.com/zapier/AutomationBench) | Business automation | Trigger data, pre-populated SaaS state, realistic tool surfaces, assertion-level scoring, an - Network accessline 25
| Benchmark | Domain | Design worth studying | |---|---|---| | [AutomationBench](https://github.com/zapier/AutomationBench) | Business automation | Trigger data, pre-populated SaaS state, realistic tool surfaces, assertion-level scoring, an - Network accessline 26
|---|---|---| | [AutomationBench](https://github.com/zapier/AutomationBench) | Business automation | Trigger data, pre-populated SaaS state, realistic tool surfaces, assertion-level scoring, and a separate strict completion result | | [AppW - Network accessline 26
|---|---|---| | [AutomationBench](https://github.com/zapier/AutomationBench) | Business automation | Trigger data, pre-populated SaaS state, realistic tool surfaces, assertion-level scoring, and a separate strict completion result | | [AppW - Network accessline 27
| [AutomationBench](https://github.com/zapier/AutomationBench) | Business automation | Trigger data, pre-populated SaaS state, realistic tool surfaces, assertion-level scoring, and a separate strict completion result | - Network accessline 27
| [AutomationBench](https://github.com/zapier/AutomationBench) | Business automation | Trigger data, pre-populated SaaS state, realistic tool surfaces, assertion-level scoring, and a separate strict completion result | - Network accessline 28
| [AppWorld](https://github.com/StonyBrookNLP/appworld) | Cross-application APIs | Stateful simulated applications, database-backed truth, cross-app workflows, and final-state verification | - Network accessline 28
| [AppWorld](https://github.com/StonyBrookNLP/appworld) | Cross-application APIs | Stateful simulated applications, database-backed truth, cross-app workflows, and final-state verification | - Network accessline 29
| [ToolSandbox](https://github.com/apple/ToolSandbox) | Stateful tool use | Controlled hidden state, milestones, tool-call consequences, and reproducible interaction state | - Network accessline 29
| [ToolSandbox](https://github.com/apple/ToolSandbox) | Stateful tool use | Controlled hidden state, milestones, tool-call consequences, and reproducible interaction state | - Network accessline 30
| [TheAgentCompany](https://github.com/TheAgentCompany/TheAgentCompany) | Workplace agents | Longer work across applications, documents, messages, and durable artifacts | - Network accessline 30
| [TheAgentCompany](https://github.com/TheAgentCompany/TheAgentCompany) | Workplace agents | Longer work across applications, documents, messages, and durable artifacts | - Network accessline 31
| [WebArena](https://github.com/web-arena-x/webarena) | Browser agents | Reproducible websites, browser interaction, seeded application state, and state-based evaluation | - Network accessline 31
| [WebArena](https://github.com/web-arena-x/webarena) | Browser agents | Reproducible websites, browser interaction, seeded application state, and state-based evaluation | - Network accessline 32
| [WorkArena](https://github.com/ServiceNow/WorkArena) | Enterprise browser work | Role-aware enterprise tasks, realistic web workflows, and application state | - Network accessline 32
| [WorkArena](https://github.com/ServiceNow/WorkArena) | Enterprise browser work | Role-aware enterprise tasks, realistic web workflows, and application state | - Network accessline 33
| [WorkArena++](https://github.com/ServiceNow/WorkArena) | Multi-application enterprise work | Longer workflows across enterprise applications and role-sensitive state | - Network accessline 33
| [WorkArena++](https://github.com/ServiceNow/WorkArena) | Multi-application enterprise work | Longer workflows across enterprise applications and role-sensitive state | - Network accessline 34
| [OSWorld](https://github.com/xlang-ai/OSWorld) | Desktop agents | OS and application snapshots, computer input, cross-application work, and state restoration | - Network accessline 34
| [OSWorld](https://github.com/xlang-ai/OSWorld) | Desktop agents | OS and application snapshots, computer input, cross-application work, and state restoration | - Network accessline 35
| [SWE-bench](https://github.com/SWE-bench/SWE-bench) | Coding agents | Real issues, pinned repository revisions, behavior tests, and patch verification | - Network accessline 35
| [SWE-bench](https://github.com/SWE-bench/SWE-bench) | Coding agents | Real issues, pinned repository revisions, behavior tests, and patch verification | - Network accessline 36
| [SWE-smith](https://github.com/SWE-bench/SWE-smith) | Scalable coding Tasks | Reproducible mutation and execution pipelines; mutations still need usefulness review | - Network accessline 36
| [SWE-smith](https://github.com/SWE-bench/SWE-smith) | Scalable coding Tasks | Reproducible mutation and execution pipelines; mutations still need usefulness review | - Network accessline 37
| [CRMArena-Pro](https://github.com/SalesforceAIResearch/CRMArena) | CRM agents | CRM roles, linked business records, permissions, and realistic sales workflows | - Network accessline 37
| [CRMArena-Pro](https://github.com/SalesforceAIResearch/CRMArena) | CRM agents | CRM roles, linked business records, permissions, and realistic sales workflows | - Network accessline 38
| [tau-bench](https://github.com/sierra-research/tau-bench) | Policy and tool use | User simulation, policy-guided decisions, tool use, and database final state | - Network accessline 38
| [tau-bench](https://github.com/sierra-research/tau-bench) | Policy and tool use | User simulation, policy-guided decisions, tool use, and database final state | - Network accessline 39
| [tau-squared Bench](https://github.com/sierra-research/tau2-bench) | Longer multi-turn work | Both sides of a tool-mediated dialogue, longer interactions, and final-state checks | - Network accessline 39
| [tau-squared Bench](https://github.com/sierra-research/tau2-bench) | Longer multi-turn work | Both sides of a tool-mediated dialogue, longer interactions, and final-state checks | - Network accessline 40
| [Terminal-Bench](https://github.com/laude-institute/terminal-bench) | Terminal agents | Isolated command-line Environments, natural instructions, tests, and reference solutions | - Network accessline 40
| [Terminal-Bench](https://github.com/laude-institute/terminal-bench) | Terminal agents | Isolated command-line Environments, natural instructions, tests, and reference solutions | - Network accessline 165
- [AutomationBench](https://github.com/zapier/AutomationBench): realistic - Network accessline 165
- [AutomationBench](https://github.com/zapier/AutomationBench): realistic - Network accessline 167
pre-populated state, tool surface, assertion-based final-state checks, assertion-level partial credit, and separate strict completion result. - [AppWorld](https://github.com/StonyBrookNLP/appworld): stateful APIs and - Network accessline 167
pre-populated state, tool surface, assertion-based final-state checks, assertion-level partial credit, and separate strict completion result. - [AppWorld](https://github.com/StonyBrookNLP/appworld): stateful APIs and - Network accessline 168
assertion-level partial credit, and separate strict completion result. - [AppWorld](https://github.com/StonyBrookNLP/appworld): stateful APIs and database verification across applications. - Network accessline 168
assertion-level partial credit, and separate strict completion result. - [AppWorld](https://github.com/StonyBrookNLP/appworld): stateful APIs and database verification across applications. - Network accessline 169
- [AppWorld](https://github.com/StonyBrookNLP/appworld): stateful APIs and - Network accessline 169
- [AppWorld](https://github.com/StonyBrookNLP/appworld): stateful APIs and - Network accessline 170
database verification across applications. - [ToolSandbox](https://github.com/apple/ToolSandbox): stateful tool calls, milestones, and controlled interaction state. - Network accessline 170
database verification across applications. - [ToolSandbox](https://github.com/apple/ToolSandbox): stateful tool calls, milestones, and controlled interaction state. - Network accessline 171
- [ToolSandbox](https://github.com/apple/ToolSandbox): stateful tool calls, - Network accessline 171
- [ToolSandbox](https://github.com/apple/ToolSandbox): stateful tool calls, - Network accessline 172
milestones, and controlled interaction state. - [TheAgentCompany](https://github.com/TheAgentCompany/TheAgentCompany): longer-running workplace workflows and artifacts. - Network accessline 172
milestones, and controlled interaction state. - [TheAgentCompany](https://github.com/TheAgentCompany/TheAgentCompany): longer-running workplace workflows and artifacts. - Network accessline 173
- [TheAgentCompany](https://github.com/TheAgentCompany/TheAgentCompany): - Network accessline 173
- [TheAgentCompany](https://github.com/TheAgentCompany/TheAgentCompany): - Network accessline 295
- [WebArena](https://github.com/web-arena-x/webarena): reproducible websites, - Network accessline 295
- [WebArena](https://github.com/web-arena-x/webarena): reproducible websites, - Network accessline 296
browser interaction, and state-based evaluation. - [WorkArena](https://github.com/ServiceNow/WorkArena): enterprise web tasks, roles, and workflow state. - Network accessline 296
browser interaction, and state-based evaluation. - [WorkArena](https://github.com/ServiceNow/WorkArena): enterprise web tasks, roles, and workflow state. - Network accessline 297
- [WorkArena](https://github.com/ServiceNow/WorkArena): enterprise web tasks, - Network accessline 297
- [WorkArena](https://github.com/ServiceNow/WorkArena): enterprise web tasks, - Network accessline 298
roles, and workflow state. - [OSWorld](https://github.com/xlang-ai/OSWorld): operating-system and desktop application snapshots, computer input, and cross-application tasks. - Network accessline 298
roles, and workflow state. - [OSWorld](https://github.com/xlang-ai/OSWorld): operating-system and desktop application snapshots, computer input, and cross-application tasks. - Network accessline 299
- [OSWorld](https://github.com/xlang-ai/OSWorld): operating-system and desktop - Network accessline 299
- [OSWorld](https://github.com/xlang-ai/OSWorld): operating-system and desktop - Network accessline 419
- [SWE-bench](https://github.com/SWE-bench/SWE-bench): real repository issues, - Network accessline 419
- [SWE-bench](https://github.com/SWE-bench/SWE-bench): real repository issues, - Network accessline 420
pinned commits, and test-based patch verification. - [SWE-smith](https://github.com/SWE-bench/SWE-smith): programmatic mutation and scalable task generation. Borrow reproducible generation and execution; do - Network accessline 420
pinned commits, and test-based patch verification. - [SWE-smith](https://github.com/SWE-bench/SWE-smith): programmatic mutation and scalable task generation. Borrow reproducible generation and execution; do - Network accessline 421
- [SWE-smith](https://github.com/SWE-bench/SWE-smith): programmatic mutation and - Network accessline 421
- [SWE-smith](https://github.com/SWE-bench/SWE-smith): programmatic mutation and - Network accessline 422
scalable task generation. Borrow reproducible generation and execution; do not assume every mutation is useful work. - [Terminal-Bench](https://github.com/laude-institute/terminal-bench): isolated - Network accessline 422
scalable task generation. Borrow reproducible generation and execution; do not assume every mutation is useful work. - [Terminal-Bench](https://github.com/laude-institute/terminal-bench): isolated - Network accessline 423
not assume every mutation is useful work. - [Terminal-Bench](https://github.com/laude-institute/terminal-bench): isolated command-line Environments and reference solutions for tasks that cross system - Network accessline 423
not assume every mutation is useful work. - [Terminal-Bench](https://github.com/laude-institute/terminal-bench): isolated command-line Environments and reference solutions for tasks that cross system - Network accessline 424
- [Terminal-Bench](https://github.com/laude-institute/terminal-bench): isolated - Network accessline 424
- [Terminal-Bench](https://github.com/laude-institute/terminal-bench): isolated - Network accessline 542
- [AppWorld](https://github.com/StonyBrookNLP/appworld): stateful application - Network accessline 542
- [AppWorld](https://github.com/StonyBrookNLP/appworld): stateful application - Network accessline 543
APIs, cross-application tasks, and database-based verification. - [CRMArena-Pro](https://github.com/SalesforceAIResearch/CRMArena): CRM roles, records, and business workflows. Use domain structure, not private data. - Network accessline 543
APIs, cross-application tasks, and database-based verification. - [CRMArena-Pro](https://github.com/SalesforceAIResearch/CRMArena): CRM roles, records, and business workflows. Use domain structure, not private data. - Network accessline 544
- [CRMArena-Pro](https://github.com/SalesforceAIResearch/CRMArena): CRM roles, - Network accessline 544
- [CRMArena-Pro](https://github.com/SalesforceAIResearch/CRMArena): CRM roles, - Network accessline 545
records, and business workflows. Use domain structure, not private data. - [WorkArena++](https://github.com/ServiceNow/WorkArena): enterprise application workflows and role-sensitive state. - Network accessline 545
records, and business workflows. Use domain structure, not private data. - [WorkArena++](https://github.com/ServiceNow/WorkArena): enterprise application workflows and role-sensitive state. - Network accessline 546
- [WorkArena++](https://github.com/ServiceNow/WorkArena): enterprise application - Network accessline 546
- [WorkArena++](https://github.com/ServiceNow/WorkArena): enterprise application - Network accessline 547
workflows and role-sensitive state. - [tau-bench](https://github.com/sierra-research/tau-bench) and [tau-squared Bench](https://github.com/sierra-research/tau2-bench): policy, - Network accessline 547
workflows and role-sensitive state. - [tau-bench](https://github.com/sierra-research/tau-bench) and [tau-squared Bench](https://github.com/sierra-research/tau2-bench): policy, - Network accessline 548
- [tau-bench](https://github.com/sierra-research/tau-bench) and - Network accessline 548
- [tau-bench](https://github.com/sierra-research/tau-bench) and - Network accessline 549
[tau-squared Bench](https://github.com/sierra-research/tau2-bench): policy, - Network accessline 549
[tau-squared Bench](https://github.com/sierra-research/tau2-bench): policy, - Network accessline 550
user interaction, tool use, and final-state checks. - [TheAgentCompany](https://github.com/TheAgentCompany/TheAgentCompany): longer workplace tasks across tools and artifacts. - Network accessline 550
user interaction, tool use, and final-state checks. - [TheAgentCompany](https://github.com/TheAgentCompany/TheAgentCompany): longer workplace tasks across tools and artifacts. - Network accessline 551
- [TheAgentCompany](https://github.com/TheAgentCompany/TheAgentCompany): longer - Network accessline 551
- [TheAgentCompany](https://github.com/TheAgentCompany/TheAgentCompany): longer - Network accessline 675
- [tau-bench](https://github.com/sierra-research/tau-bench): user simulation, - Network accessline 675
- [tau-bench](https://github.com/sierra-research/tau-bench): user simulation, - Network accessline 676
policy-guided tool use, and database final state. - [tau-squared Bench](https://github.com/sierra-research/tau2-bench): longer interactions and both sides of a tool-mediated dialogue. - Network accessline 676
policy-guided tool use, and database final state. - [tau-squared Bench](https://github.com/sierra-research/tau2-bench): longer interactions and both sides of a tool-mediated dialogue. - Network accessline 677
- [tau-squared Bench](https://github.com/sierra-research/tau2-bench): longer - Network accessline 677
- [tau-squared Bench](https://github.com/sierra-research/tau2-bench): longer - Network accessline 678
interactions and both sides of a tool-mediated dialogue. - [ToolSandbox](https://github.com/apple/ToolSandbox): stateful tool use, milestones, and controlled hidden state. - Network accessline 678
interactions and both sides of a tool-mediated dialogue. - [ToolSandbox](https://github.com/apple/ToolSandbox): stateful tool use, milestones, and controlled hidden state. - Network accessline 679
- [ToolSandbox](https://github.com/apple/ToolSandbox): stateful tool use, - Network accessline 679
- [ToolSandbox](https://github.com/apple/ToolSandbox): stateful tool use, - Network accessline 793
- [Terminal-Bench](https://github.com/laude-institute/terminal-bench): isolated - Network accessline 793
- [Terminal-Bench](https://github.com/laude-institute/terminal-bench): isolated - Network accessline 794
terminal Environments, natural instructions, tests, and reference solutions. Pin the exact 2.x or 3.x release because formats and Verifiers can change. - [SWE-bench](https://github.com/SWE-bench/SWE-bench): repository repair and - Network accessline 794
terminal Environments, natural instructions, tests, and reference solutions. Pin the exact 2.x or 3.x release because formats and Verifiers can change. - [SWE-bench](https://github.com/SWE-bench/SWE-bench): repository repair and - Network accessline 795
Pin the exact 2.x or 3.x release because formats and Verifiers can change. - [SWE-bench](https://github.com/SWE-bench/SWE-bench): repository repair and test-based correctness when the terminal Task includes code changes. - Network accessline 795
Pin the exact 2.x or 3.x release because formats and Verifiers can change. - [SWE-bench](https://github.com/SWE-bench/SWE-bench): repository repair and test-based correctness when the terminal Task includes code changes. - Network accessline 796
- [SWE-bench](https://github.com/SWE-bench/SWE-bench): repository repair and - Network accessline 796
- [SWE-bench](https://github.com/SWE-bench/SWE-bench): repository repair and
config/skills/eval-engineering/scripts/compare_tool_schemas.py
config/skills/eval-engineering/scripts/snapshot_sqlite_state.py
config/skills/langchain-dependencies/SKILL.md
- Access to secretsline 303
[Credential-related source omitted] - Access to secretsline 303
[Credential-related source omitted] - Access to secretsline 307
[Credential-related source omitted] - Access to secretsline 307
[Credential-related source omitted] - Access to secretsline 308
[Credential-related source omitted] - Access to secretsline 308
[Credential-related source omitted] - Access to secretsline 309
[Credential-related source omitted] - Access to secretsline 309
[Credential-related source omitted] - Access to secretsline 310
[Credential-related source omitted] - Access to secretsline 310
[Credential-related source omitted] - Access to secretsline 311
[Credential-related source omitted] - Access to secretsline 311
[Credential-related source omitted] - Access to secretsline 312
[Credential-related source omitted] - Access to secretsline 312
[Credential-related source omitted] - Access to secretsline 313
[Credential-related source omitted] - Access to secretsline 313
[Credential-related source omitted] - Access to secretsline 314
[Credential-related source omitted] - Access to secretsline 314
[Credential-related source omitted] - Access to secretsline 318
[Credential-related source omitted] - Access to secretsline 318
[Credential-related source omitted] - Access to secretsline 319
[Credential-related source omitted] - Access to secretsline 319
[Credential-related source omitted] - Network accessline 381
To find the current canonical import for any integration, search the integrations directory: https://python.langchain.com/docs/integrations/tools/ - Network accessline 381
To find the current canonical import for any integration, search the integrations directory: https://python.langchain.com/docs/integrations/tools/ - Network accessline 382
https://python.langchain.com/docs/integrations/tools/ - Network accessline 382
https://python.langchain.com/docs/integrations/tools/
config/skills/langchain-python-quickstart/SKILL.md
config/skills/langchain-rag/SKILL.md
- Network accessline 151
loader = WebBaseLoader("https://docs.langchain.com") - Network accessline 151
loader = WebBaseLoader("https://docs.langchain.com") - Network accessline 161
const loader = new CheerioWebBaseLoader("https://docs.langchain.com"); - Network accessline 161
const loader = new CheerioWebBaseLoader("https://docs.langchain.com"); - Network accessline 242
splits, new OpenAIEmbeddings(), { collectionName: "my-collection", url: "http://localhost:8000" } - Network accessline 242
splits, new OpenAIEmbeddings(), { collectionName: "my-collection", url: "http://localhost:8000" } - Network accessline 243
new OpenAIEmbeddings(), { collectionName: "my-collection", url: "http://localhost:8000" } ); - Network accessline 243
new OpenAIEmbeddings(), { collectionName: "my-collection", url: "http://localhost:8000" } ); - Network accessline 244
{ collectionName: "my-collection", url: "http://localhost:8000" } - Network accessline 244
{ collectionName: "my-collection", url: "http://localhost:8000" }
config/skills/langchain-typescript-quickstart/SKILL.md
config/skills/langgraph-cli/SKILL.md
- Access to secretsline 118
[Credential-related source omitted] - Access to secretsline 118
[Credential-related source omitted] - Access to secretsline 156
[Credential-related source omitted] - Access to secretsline 156
[Credential-related source omitted] - Access to secretsline 171
[Credential-related source omitted] - Access to secretsline 171
[Credential-related source omitted] - Access to secretsline 183
[Credential-related source omitted] - Access to secretsline 183
[Credential-related source omitted] - Access to secretsline 196
[Credential-related source omitted] - Access to secretsline 196
[Credential-related source omitted] - Access to secretsline 246
[Credential-related source omitted] - Access to secretsline 246
[Credential-related source omitted]
config/skills/langgraph-decision-models/SKILL.md
- Access to secretsline 30
[Credential-related source omitted] - Access to secretsline 30
[Credential-related source omitted] - Access to secretsline 38
[Credential-related source omitted] - Access to secretsline 38
[Credential-related source omitted] - Network accessline 43
[Credential-related source omitted] - Network accessline 43
[Credential-related source omitted] - Network accessline 44
[Credential-related source omitted] - Network accessline 44
[Credential-related source omitted] - Access to secretsline 44
[Credential-related source omitted] - Access to secretsline 44
[Credential-related source omitted] - Network accessline 45
base_url="https://gateway.smith.langchain.com", - Network accessline 45
base_url="https://gateway.smith.langchain.com", - Access to secretsline 49
[Credential-related source omitted] - Access to secretsline 49
[Credential-related source omitted] - Network accessline 53
[Credential-related source omitted] - Network accessline 53
[Credential-related source omitted] - Network accessline 54
[Credential-related source omitted] - Network accessline 54
[Credential-related source omitted] - Access to secretsline 54
[Credential-related source omitted] - Access to secretsline 54
[Credential-related source omitted] - Network accessline 55
base_url="https://gateway.smith.langchain.com", - Network accessline 55
base_url="https://gateway.smith.langchain.com",
config/skills/langgraph-decision-models/references/conversion-playbook.md
- Network accessline 3
A six-step audit for finding where typed classification replaces a generative call, and for proving it was worth doing. The framing of steps 1–2 is adapted from [jevify](https://github.com/ryana/jevify), a prompt for investigating what a de - Network accessline 3
A six-step audit for finding where typed classification replaces a generative call, and for proving it was worth doing. The framing of steps 1–2 is adapted from [jevify](https://github.com/ryana/jevify), a prompt for investigating what a de
config/skills/langgraph-persistence/SKILL.md
config/skills/langgraph-python-quickstart/SKILL.md
config/skills/langgraph-typescript-quickstart/SKILL.md
config/skills/managed-deep-agents/SKILL.md
- Access to secretsline 109
[Credential-related source omitted] - Access to secretsline 109
[Credential-related source omitted] - Network accessline 138
- A workspace with Managed Deep Agents public beta access, and a LangSmith API key for it. - Python and [`uv`](https://docs.astral.sh/uv/) for Python projects; Node.js and npm for TypeScript. - Network accessline 138
- A workspace with Managed Deep Agents public beta access, and a LangSmith API key for it. - Python and [`uv`](https://docs.astral.sh/uv/) for Python projects; Node.js and npm for TypeScript. - Network accessline 139
- Python and [`uv`](https://docs.astral.sh/uv/) for Python projects; Node.js and npm for TypeScript. - Network accessline 139
- Python and [`uv`](https://docs.astral.sh/uv/) for Python projects; Node.js and npm for TypeScript. - Access to secretsline 173
[Credential-related source omitted] - Access to secretsline 173
[Credential-related source omitted] - Instruction override signalline 271
> **Warning — memory is shared by every caller of the deployment, and every caller can influence it.** Never store personal data, customer data, credentials, API keys, or tokens there. Treat memory content as untrusted input: it must never - Instruction override signalline 271
> **Warning — memory is shared by every caller of the deployment, and every caller can influence it.** Never store personal data, customer data, credentials, API keys, or tokens there. Treat memory content as untrusted input: it must never - Access to secretsline 283
[Credential-related source omitted] - Access to secretsline 283
[Credential-related source omitted] - Access to secretsline 294
[Credential-related source omitted] - Access to secretsline 294
[Credential-related source omitted] - Network accessline 358
"langchainDocs": { "transport": "http", "url": "https://docs.langchain.com/mcp", - Network accessline 358
"langchainDocs": { "transport": "http", "url": "https://docs.langchain.com/mcp", - Network accessline 359
"transport": "http", "url": "https://docs.langchain.com/mcp", "include_tools": ["search_docs_by_lang_chain"], - Network accessline 359
"transport": "http", "url": "https://docs.langchain.com/mcp", "include_tools": ["search_docs_by_lang_chain"], - Network accessline 360
"url": "https://docs.langchain.com/mcp", - Network accessline 360
"url": "https://docs.langchain.com/mcp", - Network accessline 363
"notion": { "transport": "http", "url": "https://mcp.notion.com/mcp", - Network accessline 363
"notion": { "transport": "http", "url": "https://mcp.notion.com/mcp", - Network accessline 364
"transport": "http", "url": "https://mcp.notion.com/mcp", "connection": connections.get("notion", {"type": "user"}), - Network accessline 364
"transport": "http", "url": "https://mcp.notion.com/mcp", "connection": connections.get("notion", {"type": "user"}), - Network accessline 365
"url": "https://mcp.notion.com/mcp", - Network accessline 365
"url": "https://mcp.notion.com/mcp", - Network accessline 378
langchainDocs: { transport: "http", url: "https://docs.langchain.com/mcp", - Network accessline 378
langchainDocs: { transport: "http", url: "https://docs.langchain.com/mcp", - Network accessline 379
transport: "http", url: "https://docs.langchain.com/mcp", includeTools: ["search_docs_by_lang_chain"], - Network accessline 379
transport: "http", url: "https://docs.langchain.com/mcp", includeTools: ["search_docs_by_lang_chain"], - Network accessline 380
url: "https://docs.langchain.com/mcp", - Network accessline 380
url: "https://docs.langchain.com/mcp", - Network accessline 383
notion: { transport: "http", url: "https://mcp.notion.com/mcp", - Network accessline 383
notion: { transport: "http", url: "https://mcp.notion.com/mcp", - Network accessline 384
transport: "http", url: "https://mcp.notion.com/mcp", connection: connections.get("notion", { type: "user" }), - Network accessline 384
transport: "http", url: "https://mcp.notion.com/mcp", connection: connections.get("notion", { type: "user" }), - Network accessline 385
url: "https://mcp.notion.com/mcp", - Network accessline 385
url: "https://mcp.notion.com/mcp", - Network accessline 398
mda connections catalog --json mda connections create acme-api --secret-from-env ACME_API_KEY mda connections create notion --mcp https://mcp.notion.com/mcp - Network accessline 398
mda connections catalog --json mda connections create acme-api --secret-from-env ACME_API_KEY mda connections create notion --mcp https://mcp.notion.com/mcp - Network accessline 399
[Credential-related source omitted] - Network accessline 399
[Credential-related source omitted] - Access to secretsline 399
[Credential-related source omitted] - Access to secretsline 399
[Credential-related source omitted] - Network accessline 400
mda connections create notion --mcp https://mcp.notion.com/mcp - Network accessline 400
mda connections create notion --mcp https://mcp.notion.com/mcp - Access to secretsline 401
[Credential-related source omitted] - Access to secretsline 401
[Credential-related source omitted] - Network accessline 505
MDA evals are [Harbor](https://www.harborframework.com/docs/tasks) evals. `evals/` is the canonical authored dataset; put complete Harbor tasks directly under `evals/<task>/`. `.mda/evals/` is generated and must not be edited or committed. - Network accessline 505
MDA evals are [Harbor](https://www.harborframework.com/docs/tasks) evals. `evals/` is the canonical authored dataset; put complete Harbor tasks directly under `evals/<task>/`. `.mda/evals/` is generated and must not be edited or committed. - Access to secretsline 543
[Credential-related source omitted] - Access to secretsline 543
[Credential-related source omitted] - Access to secretsline 549
[Credential-related source omitted] - Access to secretsline 549
[Credential-related source omitted] - Access to secretsline 554
[Credential-related source omitted] - Access to secretsline 554
[Credential-related source omitted]
config/skills/swarm/scripts/batching.ts
config/skills/swarm/scripts/executor.ts
config/skills/swarm/scripts/filter.ts
config/skills/swarm/scripts/index.ts
config/skills/swarm/scripts/interpolate.ts
config/skills/swarm/scripts/table.ts
config/skills/swarm/scripts/types.ts
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: 286 facts
- Network access in prose of the SKILL.mdconfig/skills/deepagents-python-quickstart/SKILL.md:10Low confidence
- Network access in prose of the SKILL.mddeepagents-python-quickstart/SKILL.md:10Low confidence
- Network access in prose of the SKILL.mdconfig/skills/deepagents-typescript-quickstart/SKILL.md:10Low confidence
- Network access in prose of the SKILL.mddeepagents-typescript-quickstart/SKILL.md:10Low confidence
- Access to secrets in prose of the SKILL.md: a variable, placeholder or call, not a stored valueconfig/skills/ecosystem-primer/SKILL.md:115Low confidence
- Access to secrets in prose of the SKILL.md: a variable, placeholder or call, not a stored valueecosystem-primer/SKILL.md:115Low confidence
- Access to secrets in prose of the SKILL.md: a variable, placeholder or call, not a stored valueconfig/skills/ecosystem-primer/SKILL.md:120Low confidence
- Access to secrets in prose of the SKILL.md: a variable, placeholder or call, not a stored valueecosystem-primer/SKILL.md:120Low confidence
- Network access in prose of the SKILL.mdconfig/skills/ecosystem-primer/SKILL.md:154Low confidence
- Network access in prose of the SKILL.mdecosystem-primer/SKILL.md:154Low confidence
- Network access in prose of the SKILL.mdconfig/skills/ecosystem-primer/SKILL.md:155Low confidence
- Network access in prose of the SKILL.mdecosystem-primer/SKILL.md:155Low confidence
- Executable script in prose of a file that runsconfig/skills/eval-engineering/references/multi-turn-simulation/harbor_example.py:1Low confidence
- Executable script in prose of a file that runseval-engineering/references/multi-turn-simulation/harbor_example.py:1Low confidence
- Executable script in prose of a file that runsconfig/skills/eval-engineering/references/multi-turn-simulation/model_user.py:1Low confidence
- Executable script in prose of a file that runseval-engineering/references/multi-turn-simulation/model_user.py:1Low confidence
- Executable script in prose of a file that runsconfig/skills/eval-engineering/references/multi-turn-simulation/runner.py:1Low confidence
- Executable script in prose of a file that runseval-engineering/references/multi-turn-simulation/runner.py:1Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:25Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:25Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:26Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:26Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:27Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:27Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:28Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:28Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:29Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:29Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:30Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:30Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:31Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:31Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:32Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:32Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:33Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:33Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:34Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:34Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:35Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:35Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:36Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:36Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:37Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:37Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:38Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:38Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:39Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:39Low confidence
- Network access in prose of documentationconfig/skills/eval-engineering/references/patterns.md:40Low confidence
- Network access in prose of documentationeval-engineering/references/patterns.md:40Low confidence
and 236 more
Works with
| Agent | Status | Evidence | Install to |
|---|---|---|---|
| Claude Code | Native | Manifest .claude-plugin/marketplace.json | .claude/skills/<skill-name>/ |
| Codex | Native | Manifest config/AGENTS.md | .agents/skills/<skill-name>/ |
| Cursor | Compatible | Standard SKILL.md config/skills/deep-agents-core/SKILL.md | .cursor/skills/<skill-name>/ |
| Gemini CLI | Compatible | Standard SKILL.md config/skills/deep-agents-core/SKILL.md | .gemini/skills/<skill-name>/ |
| GitHub Copilot | Compatible | Standard SKILL.md config/skills/deep-agents-core/SKILL.md | .github/skills/<skill-name>/ |
| OpenCode | Compatible | Standard SKILL.md config/skills/deep-agents-core/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/langchain-ai-langchain-skills/)