By name: metravel-e2e-runner description: Run and debug metravel Playwright and browser smoke scenarios, use .env.e2e safely, collect trace or screenshot evidence in ignored folders, and validate real web flows without exposing secrets.
Metravel E2E Runner
Read AGENTS.md, docs/RULES.md, docs/TESTING.md, and docs/CODEX.md before running browser or Playwright flows. Load the matching feature doc from docs/features/ for the tested area.
When to use
- Playwright smoke or regression runs in
e2e/ - Browser validation for visible web UI changes
- Reproducing flaky user flows or console/runtime regressions
- Re-testing fixes that must be proven in a real browser
Execution rules
- Use
.env.e2ecredentials when present and never print secrets. - Prefer the narrowest Playwright spec or
--grepscope that proves the scenario. - For visible web UI, check browser console errors and confirm the final state with screenshot or trace evidence when useful.
- Store traces, screenshots, videos, and temporary reports only in ignored folders such as
.codex-temp/,playwright-report/, ortest-results/. - If the scenario is blocked by local server health or external instability, report the blocker and the next concrete re-run step.
Repo specifics
- Do not treat missing production-hosted media in local dev as a frontend regression by itself.
- Do not rewrite product behavior to satisfy an outdated e2e expectation; update the test when the project rule is authoritative.
- For travel performance or Lighthouse work, hand off to
$metravel-performance-analyst.
Typical commands
npm run e2enpm run check:e2e:changednpx playwright test e2e/<spec>.ts --project=chromium --workers=1