avoid-ai-writing
Skill that audits and rewrites content to remove AI writing patterns. Use it with your favorite agents including Claude Code, OpenClaw, Codex, and Hermes.
avoid-ai-writing is an agent skill that audits and rewrites text to eliminate AI writing patterns and telltale phrases. It operates in three modes: rewrite (default, with automatic second-pass verification), detect-only (flags patterns without changing text), and edit-in-place for files. The skill uses a structured catalog of 74 pattern categories and 112-entry word replacement table across three tiers to identify and fix AI-isms like significance inflation, copula avoidance, chatbot artifacts, and promotional language. It works across Claude Code, Cowork, OpenClaw, Cursor, Hermes, Codex, and other directory-aware agents, with optional voice profiles and a deterministic detection engine for CI/CD gating.
Key Features
Use Cases
- 01Cleaning AI-assisted drafts before publishing blog posts, documentation, or social media content
- 02Auditing pull request descriptions, issue comments, and GitHub discussions for AI writing tells
- 03Enforcing human-sounding prose standards in CI/CD pipelines via GitHub Actions or pre-commit hooks
- 04Detecting AI-generated content submissions without rewriting them (detect mode for moderation workflows)
- 05Rewriting marketing copy, technical documentation, or email drafts to sound more natural
- 06Preserving authentic voice in AI-assisted writing by removing LLM structural fingerprints and boilerplate
Related Skills
View moresuperpowers
An agentic skills framework & software development methodology that works.
skills
Skills for Real Engineers. Straight from my .agents directory.
skills
Public repository for Agent Skills
ponytail
Makes your AI agent think like the laziest senior dev in the room. The best code is the code you never wrote.
avoid-ai-writing — FAQ
What is avoid-ai-writing?+
avoid-ai-writing is an agent skill that audits and rewrites content to remove AI writing patterns. It identifies 74 categories of AI tells—vocabulary, structure, formatting, and chatbot artifacts—and either flags them (detect mode) or rewrites passages to eliminate them (rewrite mode).
How do I install avoid-ai-writing for Claude Code?+
Clone the repository into ~/.claude/skills/avoid-ai-writing, copy the self-contained dist/avoid-ai-writing.md file and reference it in CLAUDE.md, or install it as a plugin via '/plugin marketplace add conorbronsdon/avoid-ai-writing' then '/plugin install avoid-ai-writing@conorbronsdon-skills'.
Which AI clients and agents does this work with?+
It supports Claude Code, Claude Cowork (as plugin), OpenClaw, Cursor, Hermes, OpenAI Codex, Windsurf, Cline, GitHub Copilot, and ChatGPT Custom GPTs. Installation methods vary: directory clone, plugin install, or paste into rules/instructions field.
Does avoid-ai-writing require API keys or external services?+
No. The skill runs entirely within your AI agent using the agent's own model context. The bundled deterministic detector is a zero-dependency JavaScript engine that calls no external APIs and sends no text over the network.
Is avoid-ai-writing free to use?+
Yes, it is MIT-licensed and free for both personal and commercial use. The repository, skill files, detector engine, and GitHub Action are all open source.
What's the difference between rewrite mode and detect mode?+
Rewrite mode (default) flags AI patterns and produces a cleaned version with a second-pass audit to catch surviving tells. Detect mode only flags patterns without rewriting, useful for auditing content you don't want altered or when patterns might be intentional.
How do I install avoid-ai-writing?+
Open the source repository on GitHub and follow its README. avoid-ai-writing is a skill — MCP Agents Market links you directly to the official repo.
Is avoid-ai-writing free?+
avoid-ai-writing is an open-source project hosted on GitHub. Check the repository for its license and any usage requirements.