Diagnoses and resolves failing or flaky Playwright end-to-end tests using a systematic debugging taxonomy.
This skill provides a structured framework for Claude to identify and repair broken Playwright tests by following a rigorous engineering workflow. It guides the AI through reproduction steps, trace analysis, and categorization of failures into specific buckets like timing issues, test isolation leaks, environment mismatches, or infrastructure flakiness. By applying targeted fixes—such as implementing web-first assertions or resolving shared state issues—and verifying results through burn-in runs, it ensures that intermittent failures are permanently stabilized rather than just temporarily patched.
Key Features
01Trace-based debugging to analyze test execution and network state
020 GitHub stars
03Automated reproduction of intermittent failures using repeat-each burn-in runs
04Multi-run verification to ensure 100% test stability after modification
05Systematic categorization of failures into timing, isolation, environment, or infrastructure buckets
06Application of best-practice fixes including web-first assertions and fixture isolation
Use Cases
01Replacing brittle 'waitForTimeout' calls with reliable web-first assertions
02Resolving test isolation issues where tests interfere with each other during parallel execution
03Debugging end-to-end tests that pass locally but fail intermittently in CI pipelines
