What problem does it solve?
It helps you quickly determine whether a failing automated test is incorrect or whether the underlying code implementation is wrong, so you can fix the right layer and restore reliable CI.
Core Features & Use Cases
- Triage and diagnosis: Reproduce the failure locally and interpret expected vs actual, stack traces, and recent changes to decide what diverged.
- Correct fix path: Apply the repair process for cases where the code is wrong versus where the test/fixtures/mocks are wrong.
- Flake handling: Identify non-deterministic failures and address them by repeated runs rather than masking them with skips or time sleeps.
- Safe output format: Produce a consistent diagnosis, fix plan, and verification report so the change is reviewable and reversible.
Quick Start
Run the failing test in isolation, decide whether the contract changed or the implementation diverged, then update the test or apply the bug fix accordingly and rerun the test enough times to confirm it is stable.