What problem does it solve? Drafts of documentation, READMEs, articles, and messages often contain recognizable AI writing patterns—banned buzzwords, binary contrasts, hollow openers, inflated importance—that make prose sound generic and machine-generated. This Skill edits or audits that prose, removing the slop while preserving the author's personal voice, cadence, and edge. ## Core Features & Use Cases - Minimal-effective editing: Rewrites drafts with the smallest changes needed, cutting 14 cataloged AI patterns (binary contrasts, throat-clearing, rhetorical setups, colon reveals, synonym cycling, fake-profound kickers, em-dash overuse, and more) while protecting distinctive voice traits. - Detection mode: Audits a draft without rewriting it, reporting each named pattern with the exact quoted line and a brief suggested fix—no scores, no AI-authorship guesses. - Banned vocabulary enforcement: Removes prohibited words and filler phrases in both English and Spanish (delve, leverage, robust, streamline, "it's worth noting", etc.). - Use Case: Paste a README draft that reads like AI output and ask for an edit; receive the cleaned draft plus a short "What changed" section explaining every modification. ## Quick Start Paste your draft and ask: edit this README to remove anything that sounds like AI while keeping my voice.