ai-slop-cleaner

To clean up AI-generated code without behavior regressions, apply a deletion-first workflow that prioritizes removing dead code and duplication before consolidation, running focused regression tests and lint gates to verify preserved logic.

The best way to remove dead code and duplication is using a plan-before-edit approach that executes deletion-first refactoring, validates boundary leaks, and runs targeted verification to produce evidence-dense reports of changed files and remaining risks.

Yes, you can limit code cleanup to specific files by providing a bounded file-scope argument, which constrains the refactor scope to prevent drifting into unrelated changes while still running lint and typecheck gates.

You can review AI code cleanup proposals before applying changes by using a reviewer-only mode, allowing you to inspect suggested deletions, boundary leak fixes, and changed files without self-approving high-impact modifications in the same pass.

Regression testing works with automated code refactoring by running focused verification gates after removing abstractions and fixing boundary leaks, ensuring intended behavior is preserved while generating evidence-dense reports of remaining risks.