What problem does it solve?
You evaluate whether tests in a diff prove the code works rather than merely existing. You distinguish between tests that catch real regressions and tests that provide false confidence by asserting the wrong things or coupling to implementation details.
Core Features & Use Cases
- Untested branches in new code: Identify new conditional paths that lack corresponding tests and trace them to ensure coverage.
- False confidence tests: Flag tests that merely assert non-throwing or truthiness instead of validating behavior.
- Brittle tests: Detect tests tightly coupled to implementation details, including exact mock call counts or order dependencies, and suggest robust alternatives.
- Edge-case and error-path coverage: Highlight missing tests for error handling, fallback logic, and exceptional scenarios.
- Diff-to-test mapping: Provide actionable recommendations to align diffs with test suites and improve overall confidence.
Quick Start
Review the latest diff and clearly report any untested branches, brittle tests, and missing error-path coverage.