Does Deslop require external tools?

The standard 'quick' and 'normal' modes use internal scripts. The 'deep' thoroughness mode can optionally utilize tools like jscpd for copy-paste detection and madge for dependency analysis if they are available in your environment.

What is considered 'AI slop' by this skill?

AI slop includes artifacts like redundant debug statements (console.log), placeholder text (Lorem ipsum), excessive AI-generated comments, empty catch blocks, and unreachable 'ghost code'.

Can I run Deslop on just my recent changes?

Yes, by using the --scope=diff argument, the skill will use git to identify and scan only the files that have changed in your current branch.

Is it safe to let Deslop auto-fix my code?

Deslop uses a certainty-based hierarchy. Only 'HIGH' certainty items, which are definitely slop and safe to remove, are recommended for auto-fixing. 'MEDIUM' and 'LOW' items are flagged for manual review.

Deslop AI Code Cleanup

Name: Deslop AI Code Cleanup
Author: agent-sh

byagent-sh

•

Developer Tools

Removes AI-generated 'slop' and artifacts from codebases while preserving functionality and minimizing diff noise.

Deslop is a specialized utility designed to bridge the gap between AI-generated code and production-ready standards. It systematically identifies 'slop'—common AI artifacts such as excessive comments, leftover debug statements, placeholder text, and ghost code. By utilizing a multi-phase detection pipeline with certainty-based findings, Deslop allows developers to automate the removal of high-confidence noise while providing a structured review process for more complex structural improvements. It integrates with repo-intel to prioritize scans based on AI contribution and test coverage, making it an essential tool for maintaining repository hygiene in an AI-assisted development workflow.

Key Features

011 GitHub stars

02Smart scoping options to scan the entire repo, specific paths, or only git diffs

03Deep analysis mode integrating jscpd and madge for structural code smells

04Certainty-based detection system (High, Medium, Low) for safe automated fixes

05Structured JSON output for seamless integration with orchestrators and CI/CD

06Repo-intel integration to target AI-written files and untested code gaps

Use Cases

01Automating the removal of console.logs and debug statements before pull requests

02Pruning documentation-to-code ratios and removing 'Lorem Ipsum' placeholders

03Sanitizing a codebase after large-scale AI refactoring or code generation

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add agent-sh/deslop deslop

For use in Claude.ai and ChatGPT

Download Skill