This one rewrites academic prose to sound less like ChatGPT and more like a human researcher actually wrote it. It catches the five telltale AI patterns: hedging soup (stacking "potentially" and "arguably" until nothing means anything), formulaic transitions (furthermore, moreover, additionally), structural monotony (every paragraph the same length), abstraction fog (various studies instead of naming three specific ones), and voice erasure (passive constructions where first person belongs). The protocol is discipline aware, so it knows when passive voice is acceptable in STEM methods sections versus when humanities writing expects first person singular. Useful if you're drafting literature reviews, methodology sections, or discussion chapters and want to strip out the patterns that make reviewers suspicious.
npx skills add https://github.com/poemswe/co-researcher --skill academic-writing