Menu

#274 by GitHub stars

callstack/agent-device

Open source device layer for AI coding agents: a CLI, MCP server, and typed Node.js API to control mobile, web, TV, and desktop apps on devices and simulators, locally or in the cloud.

CodexTestingEngineering

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

skills/agent-device/SKILL.md

Week of Nov 2, 2026 · Line counts unavailable

Before Not checked yet → After Instructions only

First observation → e6ac165961de

No earlier version was observed for comparison.

Compare on GitHub

skills/android-emulator/SKILL.md

Week of Nov 2, 2026 · Line counts unavailable

Before Not checked yet → After Instructions only

First observation → e6ac165961de

No earlier version was observed for comparison.

Compare on GitHub

skills/dogfood/SKILL.md

Week of Nov 2, 2026 · Line counts unavailable

Before Not checked yet → After Network or broad shell

First observation → e6ac165961de

No earlier version was observed for comparison.

Compare on GitHub

skills/ios-simulator/SKILL.md

Week of Nov 2, 2026 · Line counts unavailable

Before Not checked yet → After Instructions only

First observation → e6ac165961de

No earlier version was observed for comparison.

Compare on GitHub

View on GitHubAdd to my skills

Verified commit e6ac165961def2f7b80336d2c04586a3ace32ce0

git clone --filter=blob:none --no-checkout https://github.com/callstack/agent-device.git .hotskills/callstack-agent-device && git -C .hotskills/callstack-agent-device checkout --detach e6ac165961def2f7b80336d2c04586a3ace32ce0 # Copy each selected skill directory from .hotskills/callstack-agent-device into .agents/skills/<skill-name>/
5.3kGitHub stars
81stars in 7 days
4skills in this repository
669commits in 30 days

Heat

7.0
  • Star growth, last 7 daysAhead of 89% of tracked repositories
  • Star growth, last 24 hoursAhead of 84% of tracked repositories
  • Rank movement, last 7 daysAhead of 43% of tracked repositories
  • Commits, last 30 daysAhead of 99% 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

Aug 2, 2026Nov 6, 20265,282

Individual skills in this repository

How it works

agent-device

When your agent uses it

Automates Apple-platform apps (iOS, tvOS, macOS), Android devices (including Android TV and Fire TV on Fire OS), and Amazon Vega OS TV apps in Vega Virtual Devices. Use when navigating apps, taking snapshots/screenshots where supported, driving TV remotes, tapping, typing, scrolling, extracting UI info, collecting evidence, or planning agent-device CLI commands.

What it loads into context

Frontmatter

83 tokens

Always available to the agent

SKILL.md

522 tokens

When the skill is activated

Reference documents

0 tokens

Loaded on request

Token counts are estimates. Scripts run separately; they are not counted as reference context.

The playbook

    Files

    • SKILL.mdcore2,191 bytes
    • SKILL.mdcore2,191 bytes
    Read SKILL.md
    
# agent-device

For a normal app-driving task, start immediately. Do not probe first with `--help`, `--version`, `devices`, `appstate`, `snapshot`, or `screenshot`:

```bash
agent-device open <app> --foreground
```

That starts the session and returns the initial interactive snapshot with `@refs`.

Loop: act with `press|click|fill|longpress <target> ... --settle`, `scroll <direction> --settle`, or `back --settle`; continue from the printed diff, verify the named expectation (`wait text "..."`, `is`, `get`, or `find`), then run `agent-device close`, or `agent-device close --shutdown` if your session booted a local simulator or emulator, unless the user asks to keep it running.

Reaching an off-screen target is one command, not a scroll-and-check loop: `scroll down --until <selector>` scrolls until that element is on screen, and `scroll bottom` runs to the end of the content. Repeated bare `scroll down` calls are the slow way to find something.
    View the complete source on GitHub

    Skills included

    • agent-device
    • android-emulator
    • dogfood
    • ios-simulator

    Repository facts

    License
    MIT
    Created
    Jan 30, 2026
    Last pushed
    Nov 5, 2026
    Forks
    382
    Open issues
    90
    Project activity
    Active

    Rank over the last 12 weeks

    Now #274 · best #227 · down 32 in 30 days

    1. #267 on Oct 31
    2. #268 from Nov 1 to Nov 2
    3. #269 on Nov 3
    4. #270 on Nov 4
    5. #271 on Nov 5
    6. #274 since Nov 6
    62 earlier periods
    1. #226 on Aug 2
    2. #226 on Aug 7
    3. #227 on Aug 8
    4. #224 on Aug 9
    5. #230 on Aug 10
    6. #229 on Aug 11
    7. #226 on Aug 12
    8. #229 from Aug 13 to Aug 14
    9. #228 on Aug 15
    10. #227 from Aug 16 to Aug 17
    11. #231 from Aug 18 to Aug 19
    12. #229 on Aug 20
    13. #232 on Aug 21
    14. #229 on Aug 22
    15. #234 from Aug 23 to Aug 24
    16. #235 from Aug 25 to Aug 26
    17. #236 on Aug 27
    18. #234 from Aug 28 to Aug 29
    19. #232 from Aug 30 to Aug 31
    20. #237 on Sep 1
    21. #232 on Sep 2
    22. #239 on Sep 3
    23. #234 on Sep 4
    24. #237 on Sep 5
    25. #234 on Sep 6
    26. #236 from Sep 7 to Sep 8
    27. #237 on Sep 9
    28. #236 on Sep 10
    29. #237 from Sep 11 to Sep 12
    30. #239 from Sep 13 to Sep 15
    31. #237 on Sep 16
    32. #235 on Sep 17
    33. #239 on Sep 18
    34. #236 on Sep 19
    35. #239 from Sep 20 to Sep 22
    36. #240 from Sep 23 to Sep 24
    37. #239 from Sep 25 to Sep 27
    38. #241 from Sep 28 to Sep 29
    39. #238 on Sep 30
    40. #240 on Oct 1
    41. #242 on Oct 2
    42. #237 from Oct 3 to Oct 4
    43. #239 on Oct 5
    44. #240 on Oct 6
    45. #242 from Oct 7 to Oct 8
    46. #246 on Oct 9
    47. #242 on Oct 10
    48. #245 on Oct 11
    49. #249 on Oct 12
    50. #250 on Oct 13
    51. #248 on Oct 14
    52. #247 from Oct 15 to Oct 16
    53. #250 from Oct 17 to Oct 18
    54. #252 on Oct 19
    55. #253 on Oct 20
    56. #254 from Oct 21 to Oct 22
    57. #252 on Oct 23
    58. #255 on Oct 24
    59. #258 on Oct 25
    60. #260 on Oct 28
    61. #264 on Oct 29
    62. #266 on Oct 30

    Before you install

    Network or broad shell

    Automated signals, not a guarantee

    Checked Nov 6, 2026

    Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.

    Show 10 notes
    • skills/agent-device/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
    • skills/android-emulator/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
    • skills/android-emulator/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
    • skills/android-emulator/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
    • skills/dogfood/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
    • skills/dogfood/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
    • skills/dogfood/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
    • skills/ios-simulator/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
    • skills/ios-simulator/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
    • skills/ios-simulator/SKILL.md:1 Analysis incomplete: Referenced executable target unavailable; not analyzed
    • 1 broad permissionin 1 file: dogfood/SKILL.md:1
    • No dangerous shell commands
    • No network access
    • No hidden or obfuscated text
    • No instruction overrides
    • No executable scripts
    • No access to secrets
    • No requests to vote in Hot
    Show the signal with its line

    skills/dogfood/SKILL.md

    • Broad permissionsline 1
      Bash(agent-device:*), Bash(npx agent-device:*)

    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: 1 fact

    Works with

    Agent compatibility, from the files in this repository
    AgentStatusEvidenceInstall to
    Claude CodeCompatibleStandard SKILL.md skills/agent-device/SKILL.md.claude/skills/<skill-name>/
    CodexNativeManifest AGENTS.md.agents/skills/<skill-name>/
    CursorCompatibleStandard SKILL.md skills/agent-device/SKILL.md.cursor/skills/<skill-name>/
    Gemini CLICompatibleStandard SKILL.md skills/agent-device/SKILL.md.gemini/skills/<skill-name>/
    GitHub CopilotCompatibleStandard SKILL.md skills/agent-device/SKILL.md.github/skills/<skill-name>/
    OpenCodeCompatibleStandard SKILL.md skills/agent-device/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.

    You might also use

    microsoft/tui-test

    💻 control, record, & test any tui app / cli with popular terminals - makes the terminal fully accessible to AI

    308GitHub stars

    fullselfbrowsing/FSB

    DOM-first AI browser automation for Chrome and MCP clients.

    205GitHub stars

    iksnerd/adb-mcp

    MCP server for Android — drive emulators and real devices over adb from Claude Code, Cursor, or VS Code. 73 tools: screenshots, UI hierarchy, tap/swipe/type, logcat, device locks,…

    17GitHub stars

    Q00/ouroboros

    Agent OS: the agent gets smarter on its own. We just hold the line: Interview-gated, staged evaluation, budgeted evolution loop. MCP server, 14 runtimes: Claude Code, Codex CLI, G…

    6.3kGitHub stars

    dannote/figma-use

    Control Figma from the command line. Full read/write access for AI agents — create shapes, text, components, set styles, export images. 100+ commands.

    614GitHub stars

    ai-dashboad/flutter-skill

    AI-powered E2E testing for 10 platforms. 253 MCP tools. Zero config. Works with Claude, Cursor, Windsurf, Copilot. Test Flutter, React Native, iOS, Android, Web, Electron, Tauri, …

    400GitHub stars

    Add this badge to your README

    It updates every night and links to this page.

    callstack/agent-device: #274 on hotskills, 5.3k stars

    [![callstack/agent-device on hotskills](https://hotskills.dev/badge/callstack-agent-device.svg)](https://hotskills.dev/skill/callstack-agent-device/)