playwright-best-practices

Replace explicit waits like waitForTimeout with auto-retrying web-first assertions such as expect(locator).toBeVisible(). Ensure test isolation with fixtures, avoid shared mutable state between tests, and use trace viewer to diagnose race conditions in parallel runs.

Use page.route() to intercept network requests and fulfill them with custom JSON responses. For advanced scenarios like HAR recording, GraphQL mocking, or simulating network failures, the network interception references cover those patterns.

Both are valid; the skill includes a dedicated POM vs fixtures decision reference. Page Object Model encapsulates page interactions in classes, while fixtures handle setup and teardown—many suites combine both approaches.

Yes, integrate @axe-core/playwright by creating an AxeBuilder instance and calling analyze() on the page. You can scope analysis to specific elements, disable rules, filter by WCAG tags, and build a reusable a11y fixture for your suite.

Add tags like @smoke or @fast to test titles, then filter with npx playwright test --grep @smoke. This supports workflows like running fast tests on PRs and full suites nightly.

Common causes include missing environment variables, different base URLs, timing differences under CI parallelism, and unseeded databases. Use process.env for configuration, webServer in config to start services, and unique identifiers for test data.