Back to directory

Humanizer

Humanizer is a Markdown-based agent skill that rewrites AI-sounding text so it reads like it was written by a person, without changing the underlying claims or facts.

Humanizer

It works with any agent that supports skills, including Claude Code, Codex and Cursor, and can rewrite pasted text or an entire file while leaving code, data, frontmatter and link targets untouched; given a sample of a user's own writing, it matches that sample's rhythm, word choice and punctuation instead of a generic voice.

It checks a drafted rewrite against a numbered list of twenty-five "tells" of AI-generated prose, such as staging a point instead of stating it directly or dressing up an ordinary fact as a pivotal one, correcting the strongest and most frequent patterns first, and asks for a missing fact rather than inventing one. It shows its work as a first rewrite, a short critique, and a final version.

Python, MIT licensed.

More in Agent skills

See all 581
  • Provenance gate

    github.com

    A deterministic authorization gateway that sits between an AI agent and its tools, tracing where each parameter of a side-effect call came from, and blocking the call or asking a human when the source does not check out.

  • A markdown framework of system rules that restates Isaac Asimov's Four Laws of Robotics as a set of prioritized loss calculations for large-language-model agents to follow, published as a discovery file for agents and crawlers to read.

  • Vitruvius

    github.com

    An engineering research agent that runs a discover, read, synthesize, verify and review loop over technical questions, reading sources directly rather than inferring from titles, and publishing its own scored verifier benchmark, weaknesses included.