How does the semantic review work?

The skill uses Claude's reasoning capabilities to look beyond simple text matching, identifying logic-based issues like code that is too simple to need the comments currently attached to it.

What kind of issues can Jin Cleanser detect?

It detects 'slop' categories such as unnecessary comments, excessive error handling, unused imports, boilerplate docstrings, over-abstraction, and leftover debug code like console.log or print statements.

How do I run a scan on my latest changes?

You can trigger the skill by typing 'jin cleanser' or 'jin deslop'. If no path is provided, it automatically analyzes your uncommitted git diff or your most recent commit.

Does Jin Cleanser automatically modify my files?

No, Jin Cleanser is strictly a review tool. It identifies issues and generates a report, but it will never automatically delete or modify your source code.

Jin Cleanser

Name: Jin Cleanser
Author: cjinzy

bycjinzy

•

開発者ツール

Audits and identifies redundant patterns or 'slop' in AI-generated code to maintain high-quality standards.

Jin Cleanser is a specialized Claude Code Skill designed to audit AI-generated source code for common redundancies and inefficiencies, often referred to as 'slop.' By combining static analysis with semantic review, it scans files or recent git changes for issues like excessive error handling, unnecessary comments, dead code, and unused imports. It generates a detailed Markdown report with severity levels ranging from Info to Critical, helping developers refine AI outputs manually without the risk of unwanted automated modifications.

主な機能

01Seamless git integration to scan uncommitted diffs or recent commits automatically

02Hybrid static and semantic analysis of AI-generated code patterns

032 GitHub stars

04Detection of various slop categories including boilerplate docstrings and over-abstraction

05Comprehensive Markdown reports with categorized issues and suggested improvements

06Non-destructive review process that protects source code from unauthorized changes

ユースケース

01Cleaning up redundant comments and excessive error handling in AI-assisted pull requests

02Detecting leftover debug statements and dead code blocks before finalizing code commits

03Auditing large blocks of LLM-generated code for production readiness and readability

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cjinzy/jin-claude jin-cleanser

For use in Claude.ai and ChatGPT