agentleFS
cenconq25cenconq25/claude-code-app-studio/.claude/skills/smoke-check/SKILL.md

smoke-check

Skill40 starsChanged 7 months agoMITView on GitHub

Run the critical-path smoke gate before QA hand-off. Executes the automated suite, verifies the core user flows defined in the QA plan, and produces a PASS/FAIL report. A failed smoke check means the build is not ready for QA.

10 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 4

    Distinctive

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read tests/smoke/critical-paths.md.

    line 34
  2. Distinctive:

    If missing, fall back to the default smoke set:

    line 37
  3. Distinctive:

    Run the unit + integration test suite first.

    line 52
  4. Distinctive:

    If the unit or integration suite fails, stop here with verdict FAIL and list the failing tests.

    line 64
  5. Distinctive:

    If a smoke flow does not yet have automation, fall back to a manual walkthrough.

    line 80

“Common” is measured: kept by at least 6 in 10 of the 9 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

Smoke Check

A short, blocking gate that runs before manual QA invests time. If smoke fails, manual QA stops and the build goes back to engineering.


Phase 1: Confirm Build and Target

Parse --build (path to .ipa / .apk / app bundle) and --target (sim, emulator, or attached device). If neither is given, infer:

  • React Native: prompt to run npx react-native run-ios or run-android.
  • Flutter: prompt to run flutter run against the target.
  • iOS native: locate the most recent build under ~/Library/Developer/Xcode/DerivedData.
  • Android native: locate the most recent app-debug.apk under app/build/outputs/apk/.

Confirm with the user before launching anything.


Phase 2: Read the Smoke Spec

Read tests/smoke/critical-paths.md. This file is authored by /qa-plan and lists the smoke flows for the current sprint.

If missing, fall back to the default smoke set:

  1. Cold start within the budgeted ms.
  2. App reaches a stable home/landing screen.
  3. Sign-in (or guest path) succeeds.
  4. Primary CTA on home reaches its destination without a crash.
  5. App backgrounds and foregrounds without state loss.
  6. Sign-out completes cleanly.

Surface the spec to the user and confirm scope.


Phase 3: Run the Automated Suite

Run the unit + integration test suite first. The smoke gate is "build is shippable for manual QA", so a red automated suite is automatic FAIL.

Per framework via Bash:

  • React Native: npx jest --silent --reporters=default --bail=false
  • Flutter: flutter test --machine
  • iOS: xcodebuild test -scheme [scheme] -destination 'platform=iOS Simulator,name=[device]'
  • Android: ./gradlew testDebugUnitTest connectedAndroidTest

Capture pass/fail counts and any failure output.

If the unit or integration suite fails, stop here with verdict FAIL and list the failing tests.


Phase 4: Run E2E Smoke Flows

For each smoke flow in the spec, run the corresponding automated flow:

  • Maestro: maestro test tests/smoke/.maestro/[flow].yaml
  • Detox: detox test --configuration ios.sim.debug --testNamePattern smoke
  • XCUITest: xcodebuild test -only-testing:UITests/SmokeUITests
  • Espresso: ./gradlew connectedAndroidTest -PtestFilter=SmokeTest

Capture per-flow PASS / FAIL / TIMEOUT and the device the flow ran on.

If a smoke flow does not yet have automation, fall back to a manual walkthrough. Use AskUserQuestion per flow:

question: "Smoke flow: [name]\n[steps]"
options:
  - "PASS — flow completed without issue"
  - "FAIL — describe what broke"
  - "BLOCKED — cannot reach this flow"

Phase 5: Mobile Health Probes

While the build is on a target, capture quick mobile-health metrics:

  • Cold start time on Tier A device (if instrumented).
  • Memory after reaching home screen.
  • App size on disk.
  • Crash log presence — read the device's crash report directory and flag any crash dated within the smoke run window.
  • Frame drops during the primary navigation path (if a perf marker is available).

Compare each value to the budget in .claude/docs/technical-preferences.md. A red probe demotes the verdict to FAIL even if every smoke flow passed.


Phase 6: Render the Verdict

## Smoke Check — [build] — [date]

Target: [device + OS]
Tester: [name]

### Automated Suite
- Unit: [pass]/[total]
- Integration: [pass]/[total]
- E2E smoke: [pass]/[total]
[failures listed verbatim]

### Smoke Flows
| Flow | Result | Notes |
|------|--------|-------|
| Cold start | PASS | 1.6s on iPhone 14 |
| Sign-in | PASS |  |
| Foreground/background | FAIL | state lost on resume |

### Mobile Health
- Cold start: 1.6s (budget 2.0s) PASS
- Memory at home: 142 MB (budget 200 MB) PASS
- Crashes during run: 0
- App size on disk: 78 MB (budget 100 MB) PASS

### Verdict: PASS / PASS WITH WARNINGS / FAIL

[If not PASS, list specifically what must be fixed before re-running.]

Verdict rules:

  • PASS — every automated suite is green, every smoke flow passes, every health probe is within budget.
  • PASS WITH WARNINGS — automated suite green, all smoke flows pass, but a non-blocking health probe (e.g., memory usage trending upward but inside budget). Document and continue.
  • FAIL — any of: red automated suite, failing smoke flow, crash during the run, or a health probe out of budget.

Phase 7: Save Report and Update State

Ask before writing:

  • production/qa/smoke-[date].md — the verdict report.

Append to production/session-state/active.md:

## Smoke Check — [date]
- Target: [device]
- Verdict: [verdict]
- Failing flows: [list or none]
- Next: [/team-qa | fix listed issues then re-run /smoke-check]

Quality Gates / PASS-FAIL

This skill IS the gate. The verdict it emits is binding for downstream QA work.


Examples

Example 1 — sprint-end smoke on iOS sim: Jest 124/124, smoke flows 6/6, cold start 1.4s on iPhone 14, no crashes. Verdict: PASS. /team-qa sprint-04 may proceed.

Example 2 — pre-release smoke on physical Pixel: Jest 124/124, but Maestro flow foreground-background fails because the app crashes on resume. Verdict: FAIL. Producer is told to fix before manual QA starts.


Next Steps

  • PASS -> run /team-qa [sprint] to execute the manual cycle.
  • FAIL -> open a bug via /bug-report, fix, re-run /smoke-check.
  • PASS WITH WARNINGS -> note the warnings in the QA sign-off.

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.