# playwright

Official Playwright end-to-end testing best practices for AI coding agents.

```bash
npx rulepack add playwright
```

## What it covers

- **Locators** — prefer user-facing locators (`getByRole`, `getByText`, `getByLabel`) over CSS/XPath; chain and filter to disambiguate.
- **Assertions** — use web-first auto-retrying assertions (`await expect(locator).toBeVisible()`); avoid manual `isVisible()` checks and manual waits.
- **Test isolation** — independent storage/cookies/data per test; shared setup via `beforeEach`/`afterEach` and `storageState`.
- **Tooling & CI** — cross-browser projects, traces on first retry, lint with `no-floating-promises`, keep Playwright updated.
- **Scope** — test user-visible behaviour, not implementation; mock third-party services with `page.route()`.

## Source

Distilled from the official Playwright documentation: https://playwright.dev/docs/best-practices

## License

MIT