playwright-best-practices

Flaky Playwright tests usually stem from race conditions, improper waits, or state leaking between tests. Replace explicit waits with auto-waiting assertions, ensure test isolation through fixtures, and investigate parallel execution issues when flakiness only appears with multiple workers.

Page Object Model in Playwright encapsulates page locators and actions into reusable classes, keeping tests readable and maintainable. The skill also covers when to prefer fixtures over POM, since Playwright fixtures often handle setup and composition more idiomatically.

Playwright fixtures are generally preferred for shared setup and dependency injection, while POM works well for modeling complex page interactions. The skill includes a dedicated reference comparing both approaches to help you choose based on your test architecture.

Playwright mocks API responses using route interception, letting you fulfill requests with custom JSON, abort them to simulate failures, or record and replay HAR files. This covers REST and GraphQL endpoints plus third-party services like payments and OAuth.

Yes, Playwright supports Electron application testing and browser extension testing with dedicated APIs. The skill includes specific references for both, covering launch configuration, extension loading, and interaction patterns unique to these targets.

CI-only failures typically come from parallel worker resource contention, missing test isolation, timing differences, or environment configuration gaps. Check sharding setup, worker-scoped fixtures for data isolation, and trace viewer output from CI artifacts to diagnose the root cause.