Summary

stephenturner/skill-deslop is a Claude skill designed to identify and remove predictable AI-generated writing patterns from prose. It targets writers, editors, and researchers seeking to imbue their text with a more authentic, human voice, particularly in scientific and technical contexts, by eliminating common AI tells.

How It Works

This skill functions by analyzing text against a curated set of rules derived from extensive reference files (phrases.md, structures.md, tropes.md). It identifies and flags or rewrites formulaic sentence structures, filler phrases, vague declarations, and other "AI tells." The approach synthesizes patterns from tropes.fyi and stop-slop, adapting them for various prose contexts, including scientific writing conventions like passive voice in methods sections.

Quick Start & Requirements

Installation involves adding the .skill file to your Claude skills directory. Options include downloading a ZIP, directly downloading a .skill release, or building from source using zip -r deslop.skill SKILL.md references/ after cloning the repository. The primary requirement is access to Claude (web or desktop). No complex software dependencies are listed.

Highlighted Details

• Addresses dozens of AI writing patterns, including formulaic structures, filler phrases, and vague declaratives.
• Features a 1-10 scoring rubric across five dimensions: Directness, Rhythm, Trust, Authenticity, and Density.
• Provides "quick checks" and direct review/scoring prompts for text analysis.
• Synthesizes rules and examples from tropes.fyi and stop-slop projects.

Maintenance & Community

The provided README does not detail specific maintenance contributors, community channels (e.g., Discord, Slack), or a public roadmap.

Licensing & Compatibility

The project is released under the MIT License. This license is highly permissive, generally allowing for commercial use and integration into closed-source projects without significant restrictions.

Limitations & Caveats

As a Claude skill, its functionality is dependent on the Claude AI platform. The effectiveness relies on the skill's rule set accurately capturing AI writing tells and the AI's ability to interpret and apply these rules. Specific performance benchmarks or known failure modes are not detailed in the README.