What problem does it solve? Flutter projects often lack a coherent testing strategy: teams mix deprecated FlutterDriver code, skip golden file baselines, miss WCAG accessibility violations, and struggle to integrate Flutter test output into CI pipelines or external Python/JS harnesses. This Skill provides a four-layer testing architecture with decision trees, reference guides, and runnable scripts to set up, migrate, and automate Flutter tests correctly. ## Core Features & Use Cases - Four-Layer Test Architecture: Covers flutter_test widget tests, integration_test on-device execution, vm_service VM introspection, and golden file plus WCAG 2.2 accessibility testing with matchesGoldenFile and meetsGuideline. - External Harness Integration: Ships a Python VM service bridge (flutter_debug_bridge.py), a shell test runner, and a converter that transforms flutter test --machine JSON-Lines output into JUnit XML for CI tools. - Migration & Troubleshooting Guides: Includes a FlutterDriver to integration_test migration reference, reusable test helpers (waitForWidget, scrollUntilVisible), and documented failure patterns like missing ensureSemantics or stale goldens. - Use Case: A team migrating from FlutterDriver can follow the migration reference, copy the helper templates, run tests via test_runner.sh in CI mode, and publish JUnit results to GitHub Actions. ## Quick Start Ask the agent to set up Flutter widget and integration tests for your project using the four-layer architecture and wire the results into CI as JUnit XML.