What problem does it solve? Existing code often fails tests, lint checks, or quality gates, and manually cycling through fix-run-review is slow and error-prone. This Skill orchestrates an iterative pipeline that assesses failures, applies fixes, and verifies every completion criterion before finishing. ## Core Features & Use Cases - Criteria-driven fix loop: Parses completion criteria into a checklist, triages failures by complexity, and runs up to 5 fix-test-review iterations with code review gates. - TDD lane for uncovered behavior: Writes a failing regression test first for behavioral criteria lacking test coverage, then fixes the code against it. - Optimization and final verification: Removes dead code and debug artifacts, runs an over-engineering review, and confirms all criteria pass. - Use Case: After a refactor leaves 12 failing tests and lint errors, run polish with the criteria "all tests pass and pre-commit is clean" and let the pipeline fix, review, and verify until green. ## Quick Start Ask the AI to polish the current branch so that all failing tests pass and the pre-commit hooks run clean.
