Use before finishing broad or release-like changes, or when asked to verify the repository end to end.
Scanned 8/31/2026
Install to Claude Code
npx -y skills add hannibalevit/resumorph --skill release-testing-checklist --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Release Testing Checklist?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/hannibalevit-release-testing-checklist)More formats (shields.io, HTML) on the badges page.
---
name: release-testing-checklist
description: Use before finishing broad or release-like changes, or when asked to verify the repository end to end.
---
# Release Testing Checklist
## Workflow
1. Inspect `git status --short` and identify backend, extension, docs, and generated-output changes.
2. For backend changes, run all backend checks from `server/`.
3. For extension changes, run `npm test` and `npm run build` from `extension/`.
4. For Docker/runtime changes, run `make up`, check `curl http://localhost:8000/health`, inspect `make logs` as needed, then stop with `make down` unless the user wants it left running.
5. For privacy-sensitive changes, apply the `privacy-security-review` skill.
6. Confirm generated or local-only artifacts are not accidentally included.
7. Summarize pass/fail status concisely.
## Validation
Backend:
```bash
cd server
uv run ruff check .
uv run ruff format --check .
uv run mypy
uv run deptry .
uv run pytest
```
Extension:
```bash
cd extension
npm test
npm run build
```
Runtime:
```bash
make up
curl http://localhost:8000/health
make down
```
Do not run `make clean` for release validation.
## Common Failure Modes
- Backend checks are skipped locally on the assumption that CI will catch them — `server-ci.yml` does run the same gate, but only after a push, which is slower feedback than running it before one.
- Coverage fails the 85% gate although tests pass.
- `extension/dist/` changes are mistaken for source changes.
- Docker data is wiped during validation.
- Validation output is too noisy to act on.
## Expected Output
Report each check as pass, fail, or not run. For failures, include the specific command and actionable error summary, not a full log dump.
Is this your skill, or is something wrong with this listing? . Author removals are honored within 72 hours.
No comments yet. Be the first to comment!