Humanize · utkarshdubeygit bundle Rewrites AI-generated, stilted, or "over-polished" text so it reads like it was written by a specific human, not a chatbot. Use this whenever the user asks to "humanize" text, make something "sound less AI," "remove the AI tells," "de-AI" a draft, "fix the ChatGPT voice," or wants a piece edited so it won't get flagged as AI-written (Wikipedia drafts, essays, cover letters, articles, comments, LinkedIn posts, etc). Also use proactively when asked to "polish," "clean up," or "proofread" a draft and the draft itself contains classic LLM tells (em dashes, "delve/boasts/underscores," rule-of-three lists, "it's not just X, it's Y," inline-bold list headers, canned "plays a vital role" filler, tidy "in conclusion" wrap-ups) — flag this to the user and offer to humanize it, don't just leave the tells in place.
