agentsclimarketplace
Sign in

Three tier test

Skill nuwansamaranayake/AiGNITEClaudeAssets/three-tier-test

Enforces the AiGNITE Chrome to mirror to phone testing sequence with hard checkpoints. Tier one runs Playwright over the Expo web build. Tier two runs a recorded walkthrough on iPhone Mirroring or Android emulator. Tier three runs a manual checklist on a physical device. Each tier must pass before the next starts. Use this skill whenever the user mentions testing a mobile app, verifying a build, QA, asking if a build is ready to ship, or running a three-tier test, even if they do not name the skill by name.From its SKILL.md

Install
npx -y skills add nuwansamaranayake/AiGNITEClaudeAssets --skill three-tier-test

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.6 KB, 724 tokens by cl100k_base, as published. Nobody here has run it

three-tier-test

Purpose

Catch bugs at the layer that surfaces them cheapest. Chrome catches console errors and broken API calls. Mirror catches gesture and layout bugs. Physical device catches sensor, push, and battery bugs.

When to trigger

Trigger on these user phrases. Match loosely.

  • "test my app"
  • "verify the build"
  • "is this ready to ship"
  • "run the test suite"
  • "QA the mobile app"
  • "three-tier test"
  • "pre-flight test"

Behavior

Run the three tiers in order. Refuse to mark the build "ready" unless all three pass. Each tier writes a pass or fail line to TEST_RESULTS.md with a timestamp.

Tier one: Chrome (Playwright)

  1. Start the Expo web build (npx expo start --web).
  2. Wait for the bundler to settle.
  3. Run scripts/run-tier-one.ps1. It drives Playwright through every screen.
  4. Assert every screen renders without a console error.
  5. Assert every API call returns a 2xx status.
  6. Hit the same endpoints listed in scripts/smoke-test.sh. Reuse the probe list.
  7. Write pass or fail to TEST_RESULTS.md.

Fails this tier produces an immediate stop. Fix and re-run.

Tier two: Mirror (iPhone Mirroring or Android emulator)

  1. Build a development client (eas build --profile development --platform <ios|android>).
  2. Install to iPhone Mirroring (Mac) or to the Android emulator.
  3. Run scripts/run-tier-two.ps1. It loads a Maestro flow or a Detox flow against the running app.
  4. Capture screen video for review.
  5. Write pass or fail to TEST_RESULTS.md.

Tier three: physical device

  1. Generate a QR code for the development build.
  2. The user scans with their phone.
  3. Run scripts/run-tier-three.ps1. It prints a manual checklist.
  4. The user ticks items off. The script captures their pass or fail and writes to TEST_RESULTS.md.

The physical-device checklist covers the items only real hardware surfaces: push notifications, camera, location, deep links, biometric auth, performance at battery saver, screenshot blocking, app icon installation, and orientation lock.

See references/physical-device-checklist.md for the full list.

Integration with mobile-app-scaffold

The skill reads the endpoint list from scripts/smoke-test.sh produced by mobile-app-scaffold. Both must reference the same list. If the smoke test list changes, tier one detects the change at the next run.

Refusal

The skill refuses to write "READY" to TEST_RESULTS.md unless:

  • Tier one is a pass.
  • Tier two is a pass.
  • Tier three is a pass.
  • All three passes are within the last seventy-two hours.

Composes with

  • Runs after security-audit.
  • Runs before app-store-submission. The submission skill rejects builds without a recent three-tier pass.

References

What ships with it: 10 files

19.9 KB alongside SKILL.md, 3 of them executable

assets/

evals/

references/

scripts/

Keep looking

Skills are one crate of 326,758. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.