SKILL.md
Repo Audit
Run the validation pipeline and produce a short markdown report.
Input
- Open workspace (TikTok-AI-Agent). Optionally a "scope" (e.g. "backend only") – still run from root; you may skip
test:e2eif scope is backend-only.
Steps
- From repo root, run in order:
- npm run lint – note pass or first ~5 lines of errors. - npm run typecheck – note pass or first ~5 lines of errors. - npm run test – note pass or failing test names/messages. - npm run test:render – note pass or failures.
- Optionally
npm run test:e2eif full audit (requires Playwright browsers). - Produce a short markdown report with:
- Each step: ok or fail + summary. - Final line: Audit: pass or Audit: fail and list of failed steps.
Output
A concise report, e.g.:
## Repo audit
- `npm run lint` – ok
- `npm run typecheck` – ok
- `npm run test` – ok
- `npm run test:render` – ok
**Audit: pass**
Or, on failure, which steps failed and the first few error lines.
References
- [.cursor/commands/validate.md](.cursor/commands/validate.md) – validate command (same steps, no report format)
- [docs/testing.md](../../docs/testing.md) – test setup, env, commands