Phase completion checklist -- verifies tests, integration, state, and PR readiness before marking a phase done.
Scanned 9/9/2026
Install to Claude Code
npx -y skills add brandonavant/harness-engineering-guide --skill example-checklist --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Example Checklist?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/brandonavant-example-checklist)More formats (shields.io, HTML) on the badges page.
---
name: "phase-complete"
description: "Phase completion checklist -- verifies tests, integration, state, and PR readiness before marking a phase done."
---
# Phase Completion Checklist
<!-- WHY THIS SKILL EXISTS:
Agents under-report failures. When an agent says "Phase 3 complete," it often means
"I wrote the code and it looks right." This skill forces a mechanical verification
sequence that catches the gap between "looks right" and "actually works." Invoke it
at the end of every implementation phase BEFORE declaring the phase complete. -->
You have been asked to verify that the current implementation phase is ready to close.
Run through every section below in order. Do NOT skip steps. Record results inline.
---
## 1. Code Completeness
- [ ] All tasks listed in the phase scope (from your agent definition) are implemented.
- [ ] No TODO, FIXME, or HACK comments were left behind (unless explicitly deferred).
- [ ] No placeholder or stub implementations remain (empty function bodies, `pass` statements, `// TODO`).
- [ ] All new files have been staged. Run `git status` and verify nothing is untracked.
## 2. Tests Pass
Run the full test suite for your territory. Record the output.
**Backend:**
```bash
cd apps/backend && python -m pytest tests/ -v --tb=short
```
**Frontend:**
```bash
cd apps/frontend && npm run test
```
- [ ] All tests pass (zero failures, zero errors).
- [ ] No tests were skipped unless documented in your agent memory with a reason.
- [ ] New code has corresponding test coverage (endpoints have happy + error path tests, components have render tests).
## 3. Lint and Type Check
**Backend:**
```bash
cd apps/backend && ruff check . && ruff format --check .
```
**Frontend:**
```bash
cd apps/frontend && npx tsc --noEmit
```
- [ ] Zero lint errors.
- [ ] Zero type errors.
## 4. Build Verification
- [ ] Docker image builds successfully: `docker compose build <service>`
- [ ] Application starts without errors: `docker compose up -d && docker compose logs <service> | tail -20`
- [ ] Database migrations run cleanly: `docker compose run --rm backend alembic upgrade head`
## 5. Integration Smoke Test
<!-- WHY: Unit tests mock system boundaries. This catches the failures that only appear
when real services talk to each other. -->
```bash
./scripts/integration-smoke-test.sh
```
- [ ] All checks pass.
- [ ] If any check fails, the failure is investigated and fixed before proceeding.
## 6. API Contract Compliance
If this phase touched any API surface (endpoints, schemas, frontend API calls):
- [ ] Implementation matches `contracts/openapi.yaml` exactly.
- [ ] No undocumented endpoints or response fields were added.
- [ ] Any necessary deviations are recorded in your agent memory under "Contract Deviations."
## 7. Agent Memory Updated
Update your agent memory with:
- [ ] Current phase marked as complete.
- [ ] Completed tasks checklist filled in.
- [ ] Files modified this session listed.
- [ ] Test results recorded (pass count, any notable findings).
- [ ] Integration verification results recorded.
## 8. PR Readiness
- [ ] All changes are committed to a feature branch (not `main`).
- [ ] Commit messages are descriptive (not "wip" or "fix stuff").
- [ ] Branch is pushed to remote.
- [ ] PR is created with a summary of changes and test results.
---
## Result
After completing all sections, summarize:
- **Phase:** [name]
- **Status:** PASS / FAIL
- **Blockers:** [list any, or "None"]
- **Notes:** [anything the next phase needs to know]
If any section fails, do NOT declare the phase complete. Fix the issue first.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!