Agent skill

qa-validation-workflow

Full validation workflow for QA agent. Runs automated checks (type-check, lint, test, build) and browser testing with E2E tests. Use when validating implementation after code review.

View SKILL.md on GitHub Repository
Stars 163
Forks 31

Install this agent skill to your Project

npx add-skill https://github.com/majiayu000/claude-skill-registry/tree/main/skills/data/qa-validation-workflow

SKILL.md

Validation Workflow Skill

"Trust but verify – automated tests catch regressions, browser tests catch reality."

When to Use This Skill

Use when:

  • currentTask.status === "ready_for_qa"
  • Developer has committed changes
  • Ready to validate implementation

MANDATORY: Worktree Verification (CRITICAL - FIRST STEP)

⚠️ CRITICAL: You MUST validate from the CORRECT worktree where the code was implemented!

Step 1: Determine Which Worktree to Validate

Check the task's agent field to know which worktree contains the code:

Task Agent Worktree Path Branch Name Command to Navigate
developer ../developer-worktree developer-worktree cd ../developer-worktree
techartist ../techartist-worktree techartist-worktree cd ../techartist-worktree
qa . (current directory) main Stay in current directory

⚠️ NEVER validate from main if the task was implemented in a worktree!

Step 2: Navigate and Verify (MANDATORY - DO NOT SKIP)

bash
# 1. CHECK current location BEFORE doing anything
pwd
git branch --show-current

# 2. Navigate to the CORRECT worktree based on task.agent
# If task.agent == "developer":
cd ../developer-worktree

# If task.agent == "techartist":
cd ../techartist-worktree

# If task.agent == "qa":
# Stay in current directory (no cd needed)

# 3. VERIFY you're in the RIGHT place (MANDATORY CHECK)
pwd
git branch --show-current
# This MUST show the expected worktree branch!

# 4. ONLY THEN proceed with validation steps below

Step 3: If You Ran Tests in Wrong Worktree

Stop immediately and report the error:

  1. The validation results are INVALID
  2. Send message to PM: "ERROR: Validated wrong worktree. Re-running..."
  3. Navigate to correct worktree
  4. Re-run ALL validation steps

Remember: QA validates the agent's worktree, NOT main (unless task.agent = "qa")

Quick Start

bash
# Run full validation suite
npm run type-check && npm run lint && npm run build && npm run test && npm run test:e2e

Validation Pipeline

      [GATE: E2E tests MUST exist - CREATE if missing]
                  │
                  ▼
┌─────────────┐    ┌──────────┐    ┌──────────────┐    ┌──────────┐
│ Type Check  │───▶│   Lint   │───▶│ TEST CHECK   │───▶│  Build   │
│    (tsc)    │    │ (eslint) │    │  (Coverage)  │    │  (vite)  │
└─────────────┘    └──────────┘    └──────────────┘    └──────────┘
       │                │                   │                  │
       ▼                ▼                   ▼                  ▼
   Pass/Fail       Pass/Fail          Tests Missing?      Pass/Fail
       │                │                   │                  │
       │                │            ┌──────┴──────┐          │
       │                │            │             │          │
       │                │         Create Tests   CANNOT PASS  │
       │                │            │          WITHOUT TEST  │
       │                │            │             │          │
       │                └────────────┴─────────────┘          │
       │                                                      │
       └──────────────────────────────────────────────────────┘
                                          │
                                          ▼
                              ┌─────────────────────┐
                              │  E2E TEST EXECUTION │ ◄── MANDATORY GATE
                              │  (npm run test:e2e)  │     NO EXCEPTIONS
                              └─────────────────────┘
                                          │
                              ┌──────────┴──────────┐
                              │                     │
                         PASS                   FAIL
                          │                        │
                          ▼                        ▼
                    Update PRD            Fix either the code or the test, following test plan.

⚠️ MANDATORY RULE: If Unit/E2E tests don't exist, CREATE THEM before validation can pass.

DO NOT mark validation as PASSED without Unit/ E2E tests covering the task acceptance criteria.

Progressive Guide

Test Coverage Check (BEFORE Automated Checks)

  1. Load qa-test-creation skill

    bash
    Skill("qa-test-creation")

  2. Identify modified source files

    bash
    # Get files changed in this task
git diff --name-only HEAD~5 | grep '^src/'
# Or read from task context in current-task-qa.json

  3. For EACH modified source file, check test coverage:

    Unit Test Check:

    • Source: src/components/game/player/index.ts
    • Test must exist: src/tests/components/game/player/index.test.ts
    • If missing: BLOCK - invoke test-creator

    E2E Test Check:

    • Check if tests/e2e/{feature}-suite.spec.ts exists
    • Example: tests/e2e/gameplay-suite.spec.ts, tests/e2e/ui-suite.spec.ts
    • If missing: BLOCK - invoke test-creator

  4. COVERAGE VERIFICATION TABLE (Must pass ALL rows):

    Check Type Pattern Status Action
    Unit test exists src/tests/**/*.test.ts for each src/**/*.ts ✅/❌ If ❌: BLOCK
    E2E test exists tests/e2e/*-suite.spec.ts ✅/❌ If ❌: BLOCK
    Tests run npm run test passes ✅/❌ If ❌: Report bugs
    E2E tests run npm run test:e2e passes ✅/❌ If ❌: Report bugs

Acceptance Criteria Verification

For each acceptance criterion in current-task-qa.json (acceptanceCriteria array):

markdown
## Acceptance Criteria Verification

### Criterion 1: "Vehicle responds to WASD input"

- **Test**: Pressed W, A, S, D keys
- **Result**: ✅ PASS / ❌ FAIL
- **Notes**: Vehicle moves forward, left, backward, right correctly

### Criterion 2: "Physics simulation runs at 60Hz"

- **Test**: Checked physics debug panel
- **Result**: ✅ PASS / ❌ FAIL
- **Notes**: Physics running at target rate

Anti-Patterns

❌ DON'T:

  • Use Playwright MCP directly for validation
  • Assume automated tests are sufficient
  • Mark as passed without running E2E tests
  • Ignore console warnings/errors

✅ DO:

  • Always run E2E tests for validation
  • Verify each acceptance criterion via test output
  • Review test screenshots as evidence
  • Document any concerns in bug notes
  • Check console for errors in test output

Didn't find tool you were looking for?

Be as detailed as possible for better results