test-review

Review unit tests for flakiness by evaluating them against Farley's eight properties—fast, isolated, repeatable, self-validating, timely, readable, specific, and comprehensive—then flag interaction-only mocks and order coupling as fragility signals.

Test double misuse typically involves mocks and stubs that validate nothing, interaction-only verification without state assertions, and tautology theatre where trivial assertions pass regardless of actual logic, producing misleading regression safety signals.

Identify coverage gaps by checking for missing boundary tests, absent error-path validation, weak mutation-testing outcomes, and error-message string matching patterns that indicate fragile assertions rather than genuine behavioral coverage.

Prioritize test review fixes using a three-tier severity model: Blocking issues that break feedback loops first, Recommendations for maintainability improvements next, and Observations for non-critical findings, addressing the highest-severity findings first.

Check for test fragility signals during feature delivery when adding new tests, during refactoring to prevent regressions, and when reviewing failing CI suites, systematically verifying quality properties, error paths, boundaries, and potential flakiness sources.