Smithery is now a part of Arcade.dev! Read more in our announcement here
AI agent validates work through evidence-based verification with comprehensive checklists, automated validation, and proof-of-work documentation...
| Feature | Description | Guide |
|---|---|---|
| Universal Checklist | Comprehensive validation | Functional, technical, security, perf, docs |
| Evidence Collection | Proof of verification | Test output, screenshots, metrics |
| Automated Pipeline | CI/CD validation | Type check, lint, test, build, audit |
| Type-Specific Checks | Context-aware validation | UI, API, migration, bug fix, optimization |
| Pre-Commit Hooks | Catch issues early | lint-staged, related tests |
| Sign-Off Template | Formal completion record | Verified by, date, evidence links |
# Universal Verification Checklist
FUNCTIONAL:
[ ] All acceptance criteria met
[ ] Happy path works correctly
[ ] Edge cases handled
[ ] Error handling works
[ ] Input validation works
TECHNICAL:
[ ] Tests pass (unit, integration, E2E)
[ ] Coverage meets threshold (80%+)
[ ] No TypeScript/lint errors
[ ] Build succeeds
[ ] No security vulnerabilities
DOCUMENTATION:
[ ] Complex logic commented
[ ] API documentation updated
[ ] README updated if needed
# Automated Verification Script
npm run lint # Lint check
npm run typecheck # Type check
npm test # Unit tests
npm run build # Build verification
npm audit # Security audit
# Full verification
npm run verify # Runs all checks
# Evidence Documentation Template
## Test Evidence
$ npm test
Test Suites: 15 passed
Tests: 87 passed
Coverage: 85.2%
## Manual Verification
| Step | Expected | Actual | Status |
|------|----------|--------|--------|
| Login | Form appears | Works | Pass |
| Submit | Success msg | Shows | Pass |
## Performance Metrics
| Metric | Baseline | Current | Status |
|--------|----------|---------|--------|
| Bundle | 245 KB | 248 KB | Pass |
| API | 150ms | 145ms | Pass |
## Sign-Off
[x] All tests pass
[x] Manual testing complete
[x] Ready for review
Verified by: [Name] on [Date]
# Type-Specific Checklists
UI COMPONENT:
[ ] Renders in all browsers
[ ] Responsive mobile/tablet/desktop
[ ] Loading/error/empty states
[ ] Keyboard accessible
[ ] Matches design specs
API ENDPOINT:
[ ] Correct status codes
[ ] Response matches schema
[ ] Auth/authz enforced
[ ] Input validated
[ ] Documented in OpenAPI
BUG FIX:
[ ] Root cause identified
[ ] Fix addresses root cause
[ ] Regression test added
[ ] No new bugs introduced
| Do | Avoid |
|---|---|
| Run all tests before marking complete | Skipping verification steps |
| Document evidence of verification | Marking complete without evidence |
| Test both happy and unhappy paths | Assuming tests are enough |
| Verify on multiple browsers/devices | Ignoring edge cases |
| Check performance impact | Skipping security checks |
| Use automated verification | Merging without green CI |
| Include screenshots for UI | Forgetting mobile testing |
| Self-review before requesting | Rushing verification |
finishing-development-branches - Complete branch preparationexecuting-plans - Quality gates during executiondebugging-systematically - Verify fixes workdeveloping-test-driven - Build verification into process