ai-slop-cleaner

To clean up AI-generated code without breaking behavior, apply a regression-tests-first workflow that locks existing functionality before removing duplication, dead code, or brittle boundaries, ensuring systematic cleanup while preserving the original logic.

A regression-tests-first cleanup workflow is a structured process that locks code behavior with tests before executing bounded, reversible passes to remove AI-generated smells, verifying stability through diffs and linting after each step.

You refactor AI-assisted codebases by defining a cleanup plan, executing bounded passes for identified smells like wrappers or scaffolds, and verifying changes through tests and linting to ensure maintainability without altering behavior.

You should use a structured cleanup pass on bloated AI code when duplication, dead code, or brittle boundaries hinder maintainability, requiring a defined plan and evidence-rich verification to safely remove scaffolds.

AI code cleanup requires defining regression tests first to lock behavior before edits, ensuring that the structured cleanup passes remain reversible and verifiable through testing and diffs throughout the refactoring process.