What problem does it solve? When someone claims a bug is fixed, there is often no rigorous evidence: the regression test may have been modified, the baseline may never have been checked, or the test may pass on both versions and prove nothing. This Skill enforces a disciplined verification workflow so fix claims are backed by comparable, reproducible test evidence. ## Core Features & Use Cases - Baseline-first verification: Run the regression test against the original revision first, distinguishing genuine assertion failures from setup, import, collection, timeout, or execution errors. - Controlled comparison: Execute the identical test with identical inputs and equivalent environments on both baseline and candidate revisions, keeping the regression test unchanged. - Structured evidence reporting: Report baseline and candidate identifiers, commands, observed failure reasons, existing-test results, environment differences, and checks not performed. - Use Case: A contributor submits a patch claiming to fix issue #412. Use this Skill to confirm the regression test fails for the reported behavioral reason on the pre-fix commit and passes on the patched commit before approving the merge. ## Quick Start Verify that the proposed fix in the current branch resolves the reported bug by running the regression test against both the baseline and candidate revisions and reporting the results separately.
