quality-refactor

To reduce technical debt safely, this skill iteratively identifies and applies refactors while running the test suite after each change to verify zero-regression. It reverts failing changes automatically, retries with adjusted strategies, and records reflections to ensure the codebase remains stable throughout the process.

The best way to reduce technical debt is generating a structured plan with ordered, risk-prioritized refactor tasks. This skill scans the codebase to create JSON artifacts containing read-first files, convergence criteria, and verification commands, then executes changes incrementally via synchronous agent-driven edits with test verification.

Yes, you can apply this refactoring skill to a specific module path, feature area, or the entire codebase. By providing a scope parameter, the skill analyzes the targeted boundary for duplication, complexity, and dead code, generating discrete, test-verified tasks that can be audited and reverted incrementally.

Automated test verification runs the project's test suite immediately after each synchronous agent-driven edit. If tests fail, the change is automatically reverted to maintain zero-regression, and the system retries the refactor with an adjusted strategy while recording a reflection log and summary.

An automated refactor agent can detect and fix duplication, complexity, naming violations, dependency issues, dead code, and pattern violations. It scans the specified scope, creates ordered tasks to address these technical debt areas, and applies changes incrementally with test verification and controlled reverts.

If an automated refactor causes a test failure, the change is automatically reverted to ensure zero-regression. The system then retries the refactor with an adjusted strategy and records a reflection log and summary documenting the failure and the new approach for future iterations.