What problem does it solve?
It removes accumulated “slop” in a codebase—duplication, dead code, stale branches, and boundary violations—while keeping behavior unchanged through evidence-first cleanup.
Core Features & Use Cases
- Scope-bounded cleanup: Uses an explicit user scope or (when Ralph state is active)
deslop_changed_files_path to prevent accidental expansion into unrelated files.
- Behavior-preserving refactor discipline: Locks behavior to existing regression coverage and strengthens tests before making changes when coverage is weak.
- Ralph-integrated verification: When used in Ralph’s mandatory deslop pass, it records cleanup status and reruns focused verification (tests/build/lint/typecheck as applicable).
- Evidence-dense reporting: Produces an “AI SLOP CLEANUP REPORT” that lists smells removed, verification results, and residual risks.
Quick Start
Use this skill when you want a behavior-preserving cleanup of the current project scope after planning or during Ralph’s cleanup pass by telling the assistant: “Run /omagy:ai-slop-cleaner on my current change scope and verify with targeted regression checks.”
