unslop

Rewrites prose and code diffs to remove AI writing patterns and add human voice.

1|Updated Aug 26, 2026
One-click install
npx skills add https://github.com/edivad1999/stuc-stack --skill unslop-edivad1999
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: unslop
Source: https://github.com/edivad1999/stuc-stack/tree/main/skills/unslop
Command: npx skills add https://github.com/edivad1999/stuc-stack --skill unslop-edivad1999

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI-generated text carries recognizable tells: puffery, em dash overuse, rule-of-three structures, chatbot phrases, and vague attributions. This Skill detects and rewrites those patterns in prose, commits, PRs, and code diffs so the output reads as human-written. ## Core Features & Use Cases - Pattern detection and rewrite: Scans for 31 cataloged AI tells across content, language, style, filler, and jargon, then rewrites while preserving meaning and tone. - Voice restoration: Adds opinions, varied rhythm, specificity, and first person where appropriate so text is not sterile after cleanup. - Code diff cleanup: Before commit, cuts narrating comments, unsupported guards, dead compatibility paths, and unrelated edits riding the same commit. - Use Case: After drafting a verification report or PR description with an AI assistant, run this pass to strip phrases like "Additionally" and "showcasing", replace em dashes with periods, and name the actual Gradle task or android command that ran. ## Quick Start Review my PR description and commit message, remove any AI writing patterns, and rewrite them in a natural human voice.

Frequently Asked Questions about unslop

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I remove AI writing patterns from generated text?▼

Scan the text for known tells like puffery, em dash overuse, rule-of-three structures, and chatbot phrases, then rewrite each instance while preserving meaning. Finish with a self-audit asking what still makes the text obviously AI generated.

What are common signs of AI generated writing?▼

Common tells include words like delve, crucial, and tapestry, excessive em dashes, forced groups of three, bold labels with colons restating the line, vague attributions like "experts believe", and sycophantic openers like "Great question!".

Can this clean up code diffs before committing?▼

Yes. On a code diff it cuts narrating comments that restate the next line, unsupported guards, dead compatibility paths, unused parameters, and unrelated edits riding the same commit. Deeper comment review is handed to the /no-comments pass.

Does removing AI patterns make text sound human automatically?▼

No. Sterile, voiceless writing is also an AI tell, so the rewrite step adds opinions, varied sentence rhythm, first person where it fits, and specific concrete details like the exact Gradle task or command that ran.

When should I not use this editing approach?▼

Avoid it when strict formal or legal phrasing is required and voice would be inappropriate, or when the text is machine-generated output like logs where style rules do not apply. It targets prose, commits, PRs, and diffs.