ai-slop-cleaner

To remove AI-generated code bloat safely, lock behavior with focused regression tests first, then apply a deletion-first workflow to eliminate dead code, duplicates, and over-abstractions while preserving the original functionality.

The best way to clean up duplicated AI code is to run smell-focused passes that consolidate duplicates, improve naming, and reinforce error handling. Using a reviewer-only --review mode separates inspection from editing for safer consolidation.

You can perform a targeted code cleanup by applying the anti-slop pass to explicit file lists, changed-file scopes, or Ralph session changed files, limiting the risk and surface area to only the recently modified code.

Yes, you can inspect AI code slop without modifications by using the reviewer-only --review mode, which separates writer and reviewer responsibilities to surface dead code and over-abstractions for inspection before any deletion occurs.

Writing regression tests before deleting dead code locks the current behavior, ensuring that your deletion-first cleanup and duplicate consolidation do not introduce regressions or alter the working application's output.

No, ai-slop-cleaner does not require any external dependencies to run. It operates directly on bounded file lists or Ralph session changed files to perform behavior-preserving cleanup and test reinforcement.