What problem does it solve? AI-generated code often accumulates slop: obvious comments, dead code, over-defensive checks, needless abstractions, and oversized modules. Manually cleaning this up risks breaking behavior, and ad-hoc cleanup is unreviewable. This Skill removes slop safely by locking behavior with regression tests before touching a single line. ## Core Features & Use Cases - Behavior Lock First: Writes narrow regression tests to lock observable behavior before any cleanup begins; files without a green baseline are skipped. - 10 Slop Categories: Covers stylistic, structural, hidden-cost, and behavior-coverage issues including dead code, duplication, performance equivalences, and oversized modules over 250 LOC. - Parallel Batch Cleanup: Dispatches up to 5 files concurrently to task agents with per-file category checklists and hard constraints, then verifies with 5 quality gates and 3 review checklists. - Use Case: After merging an AI-assisted feature branch, run the skill to deslop the diff: it scopes changed files via git merge-base, locks behavior, cleans in batches, and reports a final CLEAN or REQUIRES ATTENTION status. ## Quick Start Ask the agent to remove AI slop from the current branch changes and it will lock behavior with tests, clean the files in parallel batches, and verify with quality gates.
