Drop-in writing skill for Claude Code and any LLM agent that scrubs AI tells out of drafts before they ship. Six-step pipeline, 16 structural patterns, 3-tier vocab system, channel-aware strictness, optional voice profiles, guided setup mode.
Measured, self-improving AI-text humanizer: register-aware stylometric scorer + LLM judge panel vs real human corpora. Claude Code skill (/humanizer-pro); runs on Codex & any file-aware agent.
One local MCP server wrapping five free prose linters (proselint, alex, write-good, blocklint, textstat) + AI-tell detector - for catching weak, inflated, exclusionary, or LLM-generated writing without subscription costs
A CLAUDE.md that strips AI-slop from markdown — no em dashes, no marketing fluff, no emoji bullets, sentence case headings. Anti-AI-tells discipline for any AI-assisted writing.
Add a description, image, and links to the ai-tells topic page so that developers can more easily learn about it.
To associate your repository with the ai-tells topic, visit your repo's landing page and select "manage topics."