No AI Slop

No AI Slop (github.com/petergyang/no-ai-slop) is Peter Yang's most popular skill, and the premise is easy to recognise: AI-assisted writing has a house style, and once you can see it you cannot unsee it. The skill catches 20+ patterns of it — binary contrasts (“It's not X. It's Y.”), throat-clearing openers (“Here's the thing”), faux-insight setups (“What nobody tells you”), colon reveals, dramatic fragments, importance puffery (“marks a pivotal moment”), weasel attribution (“experts agree”), synonym cycling, and fake-profound endings (“The future isn't coming. It's already here.”).

What makes it more than a banned-phrase list is the constraint it works under: remove the slop without flattening your voice. Most AI editing passes sand off humour, rhythm and personal quirks along with the clichés, leaving prose that is clean and dead. This one is built to preserve the writer while removing the tells, and it reports back exactly what it changed so you stay in control of the edit. It also checks fundamentals — lead with the point, active voice, untangle knotted sentences, prefer concrete details to abstractions.

Three ways to use it: /no-ai-slop (your writing) edits a draft; /no-ai-slop is this slop? (your writing) quotes every pattern it finds without guessing whether AI wrote it and without changing anything; and “Draft an AI slop post about (topic)” generates the most cringe-worthy slop it can as satire — a genuinely good way to learn the patterns. Install by asking your agent to add it globally from the repo, or with npx skills add petergyang/no-ai-slop --skill no-ai-slop --global --yes; it also ships as a ChatGPT plugin. Pairs naturally with the same author's Human Review skill: one catches the slop, the other lets you fix it visually. MIT licensed.