QA Phase Skill

SkillAI & models

Run QA tests at various levels (L1-L5) for features under PDCA management. Integrates with MCP tools to execute structured test plans.

Available today. Use it from your connected AI after setup.

Connect ahel once, and every AI you use reads what you have installed.

Then ask your AI: use the QA Phase Skill skill

What this skill tells your AI

The instructions your AI receives, as published by ww-w-ai/bkit-gemini in skills/qa-phase/SKILL.md and read by ahel’s review.

Structured QA testing with 5 levels from syntax to production-readiness

Commands

CommandDescriptionExample
/qa-phase [feature]Run default (L2) QA/qa-phase user-auth
/qa-phase [feature] --level L1Syntax check only/qa-phase user-auth --level L1
/qa-phase [feature] --level L3Full test suite/qa-phase user-auth --level L3
/qa-phase [feature] --level L5Production readiness/qa-phase user-auth --level L5

MCP Integration

This skill uses the bkit_qa_run MCP tool to execute QA tests.

How to Execute

  1. Parse the feature name and level from arguments
  2. Call bkit_qa_run with { feature, level }
  3. Display results with pass/fail indicators
  4. If failures found, suggest specific fixes

QA Levels

L1: Syntax and Lint

  • File syntax validation (node -c, tsc --noEmit, etc.)
  • Linting rules (ESLint, Prettier)
  • Import/require resolution check
  • No build errors

L2: Unit and Logic (Default)

All of L1, plus:

  • Unit test execution (Jest, Mocha, Vitest)
  • Test coverage check (minimum threshold)
  • Logic validation against Design document
  • Edge case identification

L3: Integration

All of L2, plus:

  • API endpoint testing
  • Database query validation
  • Third-party service mock testing
  • Cross-module interaction tests

L4: End-to-End

All of L3, plus:

  • User flow testing (Playwright, Cypress)
  • Browser compatibility (if web)
  • Performance benchmarks
  • Accessibility checks (a11y)

L5: Production Readiness

All of L4, plus:

  • Security scan (dependency audit, OWASP checks)
  • Load testing results review
  • Rollback verification
  • Documentation completeness
  • Deploy checklist validation

Test Plan Generation

When no existing tests are found:

  1. Read the Design document for the feature
  2. Extract testable requirements
  3. Generate a test plan with:
    • Test case descriptions
    • Expected inputs and outputs
    • Edge cases to cover
  4. Write test plan to docs/qa/{feature}-test-plan.md

Output Format

## QA Report: {feature} (Level {level})

### Summary
- Level: L2 (Unit and Logic)
- Total checks: 24
- Passed: 22
- Failed: 2
- Coverage: 87%

### Results

#### L1: Syntax and Lint
- [PASS] File syntax valid (8/8 files)
- [PASS] ESLint clean (0 errors, 2 warnings)

#### L2: Unit and Logic
- [PASS] Unit tests (18/20 passing)
- [FAIL] Missing test for edge case: empty input
- [FAIL] Coverage below threshold (87% < 90%)

### Recommendations
1. Add test for empty input handling in src/validators/user.js
2. Add tests for error paths in src/api/auth.js to reach 90% coverage

Integration with PDCA

QA Phase maps to the Check phase:

  1. Complete implementation (Do phase)
  2. Run /qa-phase {feature} --level L2 for initial check
  3. Fix issues and re-run until passing
  4. Run /qa-phase {feature} --level L3+ for deeper validation
  5. Results feed into /pdca analyze match rate calculation

Signals

GitHub stars
65
Forks
16
Last commit
May 2026
Advanced
Catalog kind
skill
Gateway key
qa-phase
Source
github.com/ww-w-ai/bkit-gemini