agentleFS
gronxbgronxb/codex-relay/.agents/skills/dogfood/SKILL.md

dogfood

Skill733 starsChanged 2 months agoApache-2.0View on GitHub

Systematically explore and test a mobile app on iOS/Android with agent-device to find bugs, UX issues, and other problems. Use when asked to "dogfood", "QA", "exploratory test", "find issues", "bug hunt", or "test this app" on mobile. Produces a structured report with reproducible evidence: screenshots, optional repro videos, and detailed steps for every issue.

81 other projects use a version of this, 14 of them edited.

Adopted by

Most recently taken up by

Used with

Projects with it also use dep-refs, feature-boundary-planner and 9 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • betacatslingSkill

    dep-refs

    Use when needs to inspect TypeScript export references in the z-code workspace, list exports from a file, verify whether an export is unused before deletion, investigate who imports a symbol during refactors, or combine pnpm knip unused-export results with pnpm dep:refs symbol-level reference tracing.

    Also in 5 projects that use this

  • betacatslingSkill

    feature-boundary-planner

    Map a ZCode behavior change to current UI surfaces, state owners, protocol commands, persistence, and validation. Use for impact analysis, product-boundary planning, or an implementation handoff grounded in the checked-out source.

    Also in 5 projects that use this

  • lir-navcooSkill

    yuanbao

    Yuanbao (元宝) groups: @mention users, query info/members.

    Also in 18 projects that use this

  • Cartograf666Skill

    architecture-governance

    Bootstrap, document, enforce, and review a project's architecture through layer responsibilities, dependency direction, architecture manifests, legacy baselines, Change Envelopes, cross-layer approval policy, and deterministic architecture checks. Use when starting a project, auditing an existing codebase, planning cross-layer work, preventing architecture erosion, reviewing structural changes, or deciding whether implementation may expand into another layer.

    Also in 5 projects that use this

  • john-data-chenSkill

    touchdesigner-mcp

    Control a running TouchDesigner instance via twozero MCP — create operators, set parameters, wire connections, execute Python, build real-time visuals. 36 native tools.

    Also in 7 projects that use this

  • math-incSkill

    minecraft-modpack-server

    Set up a modded Minecraft server from a CurseForge/Modrinth server pack zip. Covers NeoForge/Forge install, Java version, JVM tuning, firewall, LAN config, backups, and launch scripts.

    Also in 5 projects that use this

The file

Show its rules and text

We split this skill into the rules it gives an agent

16 rules. 0 are common, 10 are distinctive, and 6 are very specific to this project.

  • 0

    Common

    Most of the 81 edited versions we read keep these.

  • 10

    Distinctive

    This author's own choices. Worth a look.

  • 6

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Only the Target app is required.

    line 13
  2. Distinctive:

    If the user gives enough context to start, begin immediately with defaults.

    line 24
  3. Very specific:

    Prefer direct agent-device binary when available.

    line 26
  4. Distinctive:

    Start a named session and launch target app:

    line 48
  5. Distinctive:

    If login is required:

    line 55
  6. Distinctive:

    Read references/issue-taxonomy.md for severity/category calibration.

    line 81
  7. Distinctive:

    Test core journeys end-to-end (create, edit, delete, submit, recover).

    line 86
  8. Very specific:

    Use diff snapshot -i after UI transitions to avoid stale refs.

    line 88
  9. Distinctive:

    Use video + step screenshots:

    line 106
  10. Very specific:

    Use refs (@eN) for fast exploration, selectors for deterministic replay assertions when needed.

    line 165
  11. Very specific:

    Use fill for clear-then-type semantics.

    line 167
  12. Distinctive:

    Keep logs optional and targeted: enable/read app logs only when useful for diagnosis.

    line 168
  13. Very specific:

    If the issue appears rooted in React Native internals rather than device/app runtime behavior, use agent-device react-devtools ...

    line 169
  14. Distinctive:

    Never read source code of the app under test.

    line 170
  15. Distinctive:

    Write each issue immediately to avoid losing evidence.

    line 171
  16. Very specific:

    Never delete screenshots/videos/report artifacts during a session.

    line 172

“Common” is measured: kept by at least 6 in 10 of the 81 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Dogfood (agent-device)

Systematically explore a mobile app, find issues, and produce a report with full reproduction evidence for every finding.

Setup

Only the Target app is required. Everything else has sensible defaults.

ParameterDefaultExample override
Target app(required)Settings, com.example.app, deep link URL
PlatformInfer from user context; otherwise ask (ios or android)--platform ios
Session nameSlugified app/platform (for example settings-ios)--session my-session
Output directory./dogfood-output/Output directory: /tmp/mobile-qa
ScopeFull appFocus on onboarding and profile
AuthenticationNoneSign in to user@example.com

If the user gives enough context to start, begin immediately with defaults. Ask follow-up only when a required detail is missing (for example platform or credentials).

Prefer direct agent-device binary when available.

Workflow

1. Initialize    Set up session, output dirs, report file
2. Launch/Auth   Open app and sign in if needed
3. Orient        Capture initial snapshot and map navigation
4. Explore       Systematically test flows and states
5. Document      Record reproducible evidence per issue
6. Wrap up       Reconcile summary, close session

1. Initialize

mkdir -p {OUTPUT_DIR}/screenshots {OUTPUT_DIR}/videos
cp {SKILL_DIR}/templates/dogfood-report-template.md {OUTPUT_DIR}/report.md

2. Launch/Auth

Start a named session and launch target app:

agent-device --session {SESSION} open {TARGET_APP} --platform {PLATFORM}
agent-device --session {SESSION} snapshot -i

If login is required:

agent-device --session {SESSION} snapshot -i
agent-device --session {SESSION} fill @e1 "{EMAIL}"
agent-device --session {SESSION} fill @e2 "{PASSWORD}"
agent-device --session {SESSION} press @e3
agent-device --session {SESSION} wait 1000
agent-device --session {SESSION} snapshot -i

For OTP/email codes: ask the user, wait for input, then continue.

3. Orient

Capture initial evidence and navigation anchors:

agent-device --session {SESSION} screenshot {OUTPUT_DIR}/screenshots/initial.png
agent-device --session {SESSION} snapshot -i

Map top-level navigation, tabs, and key workflows before deep testing.

4. Explore

Read references/issue-taxonomy.md for severity/category calibration.

Strategy:

  • Move through each major app area (tabs, drawers, settings pages).
  • Test core journeys end-to-end (create, edit, delete, submit, recover).
  • Validate edge states (empty/error/loading/offline/permissions denied).
  • Use diff snapshot -i after UI transitions to avoid stale refs.
  • Periodically capture logs path and inspect the app log when behavior looks suspicious.

Useful commands per screen:

agent-device --session {SESSION} snapshot -i
agent-device --session {SESSION} screenshot {OUTPUT_DIR}/screenshots/{screen-name}.png
agent-device --session {SESSION} appstate
agent-device --session {SESSION} logs path

5. Document Issues (Repro-First)

Explore and document in one pass. When you find an issue, stop and fully capture evidence before continuing.

Interactive/behavioral issues

Use video + step screenshots:

  1. Start recording:
agent-device --session {SESSION} record start {OUTPUT_DIR}/videos/issue-{NNN}-repro.mp4
  1. Reproduce with visible pacing. Capture each step:
agent-device --session {SESSION} screenshot {OUTPUT_DIR}/screenshots/issue-{NNN}-step-1.png
sleep 1
# perform action
sleep 1
agent-device --session {SESSION} screenshot {OUTPUT_DIR}/screenshots/issue-{NNN}-step-2.png
  1. Capture final broken state:
sleep 2
agent-device --session {SESSION} screenshot {OUTPUT_DIR}/screenshots/issue-{NNN}-result.png
  1. Stop recording:
agent-device --session {SESSION} record stop
  1. Append issue immediately to report with numbered steps and screenshot references.

Static/on-load issues

Single screenshot is sufficient; no video required:

agent-device --session {SESSION} screenshot {OUTPUT_DIR}/screenshots/issue-{NNN}.png

Set Repro Video to N/A in the report.

6. Wrap Up

Target 5-10 well-evidenced issues, then finish:

  1. Reconcile summary severity counts in report.md.
  2. Close session:
agent-device --session {SESSION} close
  1. Report total issues, severity breakdown, and highest-risk findings.

Guidance

  • Repro quality matters more than issue count.
  • Use refs (@eN) for fast exploration, selectors for deterministic replay assertions when needed.
  • Re-snapshot after any mutation (navigation, modal, list update, form submit).
  • Use fill for clear-then-type semantics; use type for incremental typing behavior checks.
  • Keep logs optional and targeted: enable/read app logs only when useful for diagnosis.
  • If the issue appears rooted in React Native internals rather than device/app runtime behavior, use agent-device react-devtools ... and the skills/react-devtools workflow for component-tree or render-profiling inspection.
  • Never read source code of the app under test; findings must come from observed runtime behavior.
  • Write each issue immediately to avoid losing evidence.
  • Never delete screenshots/videos/report artifacts during a session.

References

ReferenceWhen to Read
references/issue-taxonomy.mdStart of session; severity/categories/checklist

Templates

TemplatePurpose
templates/dogfood-report-template.mdCopy into output directory as the report file

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.