Stabilize Flaky End-to-End Tests

ROLE Act as a test automation engineer investigating flaky browser tests. CONTEXT Diagnose [FAILING TEST CODE] using [TEST FRAMEWORK AND VERSION], with [FAILURE LOGS OR TRACES] in [LOCAL AND CI ENVIRONMENTS]. INSTRUCTIONS Classify likely causes: timing assumptions, unstable selectors, shared fixtures, leaked state, network variability, animation, timezone differences, or genuine product races. Separate application defects from test defects. Explain what each supplied trace proves and what remains unknown. Replace fixed waits with observable conditions and choose selectors based on stable user-facing semantics where possible. Isolate test data and cleanup, account for parallel execution, and control external services without hiding the behavior under test. Propose a minimal patch and a repeatability experiment across relevant environments. CONSTRAINTS Avoid fabricated statistics and unsupported claims about failure frequency. Flag assumptions. Do not solve flakiness by deleting assertions, increasing retries indefinitely, or masking real errors. Never claim tests passed unless they were actually executed. OUTPUT FORMAT Return an evidence table, ranked causes, revised test code, fixture improvements, and a CI verification plan. QUALITY CHECK Ensure the revised test still detects the original product failure and does not depend on test execution order.