ai-slop-cleaner

To clean up AI-generated code slop safely, you need a regression-tests-first workflow that performs smell-by-smell passes on a bounded set of files, ensuring minimal diffs and behavior preservation before committing changes.

A bounded scope cleanup workflow restricts refactoring to a defined set of changed files or feature areas, categorizing issues like dead code and duplication into structured passes to avoid broad, uncontrolled rewrites.

You refactor code smells one pass at a time by categorizing issues into duplication, dead code, boundary violations, and naming clarity, then addressing each category sequentially while running lint and type checks.

Yes, explicit regression tests are required before cleaning up dead code and duplication to verify that the refactor preserves the original behavior and produces an audit-ready report of remaining risks.

The best way to reduce AI slop in scoped edits is to integrate a changed-files workflow that requires a bounded cleanup plan, explicit verification, and minimal diffs before any code quality improvements are committed.

You should not use a bounded cleanup workflow when you lack explicit regression tests or when the refactoring task requires a broad rewrite that exceeds the defined set of changed files in your feature area.