Can I automatically fix the detected issues?

Yes, the skill generates a report with specific suggestions and can be used alongside the scribe:doc-generator skill with a remediation flag to apply fixes.

What exactly is 'AI slop'?

AI slop refers to the repetitive, predictable, and often vapid linguistic patterns frequently found in AI-generated text, such as the overuse of specific buzzwords or overly formal transition phrases.

Is it okay to use some AI-common words occasionally?

Yes. The detector focuses on density and clustering rather than individual word usage, acknowledging that some terms are perfectly valid in specific contexts.

How does the Slop Density Score work?

It calculates a score based on the frequency of Tier 1 (high-confidence) and Tier 2 (medium-confidence) markers relative to the total word count, rating content from 'Clean' to 'Heavy' slop.

Does this skill work on source code?

The skill is optimized for prose, but it effectively scans markdown files, text files, and docstrings within code to ensure technical explanations are concise.

AI Slop Detector

Name: AI Slop Detector
Author: athola

byathola

•

130

•

コンテンツ管理

Identifies and remediates AI-generated linguistic markers and patterns in documentation to improve content quality and authenticity.

The AI Slop Detector is a specialized tool for Claude Code designed to audit documentation, technical prose, and blog posts for 'AI slop'—the repetitive vocabulary, predictable structural patterns, and sycophantic phrasing common in LLM-generated text. By analyzing vocabulary density, em dash usage, and sentence length uniformity, it provides a quantitative 'Slop Density Score' and actionable remediation reports. This skill is essential for teams who use AI to draft content but want to ensure their final public-facing documentation maintains a high-quality, human-centric voice.

主な機能

01Automated Slop Density Score calculation to categorize content quality

02Multi-tier vocabulary scanning for high-confidence AI markers like 'delve' and 'tapestry'

03130 GitHub stars

04Structural analysis of em dash density, tricolon patterns, and sentence length variance

05Detection of sycophantic conversational fillers and empty marketing jargon

06Detailed reporting with specific replacement suggestions for identified markers

ユースケース

01Auditing project READMEs and documentation to remove robotic or repetitive phrasing

02Cleaning up draft content generated by LLMs to ensure a professional, human-like tone

03Reviewing technical blog posts and tutorials for clarity and impact before publication

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add athola/claude-night-market slop-detector

For use in Claude.ai and ChatGPT

Download Skill