/qa — QA Lead

Find bugs before users do. Every fix gets a regression test.

When to Use

• After implementing a feature
• After /review flags issues
• Before /ship
• When you suspect edge cases

Process

Step 1: Detect Test Framework

Auto-detect what's available:

• package.json → Jest, Vitest, Mocha, Playwright
• pytest.ini / pyproject.toml → pytest
• go.mod → go test
• Cargo.toml → cargo test
• Nothing? → Bootstrap one (ask user which)

Step 2: Run Existing Tests

# Run the test suite, capture output
{test_command} 2>&1

Report: X passed, Y failed, Z skipped. Coverage: N%.

Step 3: Analyze for Gaps

Read the diff and identify:

• New functions without tests
• Changed logic without updated tests
• Edge cases: null, empty, boundary values, concurrent access
• Error paths: what happens when things fail?

Step 4: Write Missing Tests

For each gap, write a test that:

• Has a clear name describing the scenario
• Tests ONE thing
• Includes the edge case that would catch a regression
• Follows the project's existing test patterns

Step 5: Run Again

Run the full suite including new tests. If anything fails:

1. Diagnose the failure
2. Fix the code (not the test, unless the test is wrong)
3. Commit the fix atomically
4. Re-run to verify

Step 6: Report

## QA Report

Tests: 42 → 51 (+9 new)
Coverage: 83% → 87% (+4%)
Bugs found: 2
Bugs fixed: 2

### New Tests
- test_notification_empty_payload (edge case)
- test_notification_duplicate_delivery (race condition)
- ...

### Bugs Fixed
- 🐛 Empty payload caused crash → now returns 400
- 🐛 Duplicate delivery on retry → added idempotency key

Chaining

• Before: /review (review findings inform what to test)
• After: /ship (won't ship if tests fail)