Avoid-ai-writing audits and rewrites text to remove AI writing patterns, supporting detect-only and edit-in-place modes across Claude Code, OpenClaw, Hermes, and other agents. Latest releases add a style option for user-supplied house-style configuration.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
Avoid-ai-writing is a JavaScript-based skill that audits and rewrites content to remove AI writing patterns. It supports detect-only and edit-in-place modes, plus voice profiles. It is compatible with Claude Code, OpenClaw, Hermes, and other agents.
How it works
The tool performs audits to identify AI writing patterns and can rewrite text to remove them (Rewrite mode), or flag patterns without rewriting (Detect mode). It includes a two-pass approach: an initial edit followed by a second pass to catch surviving patterns. It also offers an optional voice profile setting and a configurable house-style layer when used with the new --style option.
Getting started
Installation & Usage details are provided for multiple targets:
- Claude Code: clone into skills directory, copy the file, or use as a slash command. Example commands:
git clone https://github.com/conorbronsdon/avoid-ai-writing ~/.claude/skills/avoid-ai-writing
- OpenClaw: install from ClawHub or clone into skills directory:
clawhub install avoid-ai-writing
git clone https://github.com/conorbronsdon/avoid-ai-writing ~/.openclaw/skills/avoid-ai-writing
- Cursor: place the ported rule in your project’s .cursor/rules/ with an example curl command:
curl -o .cursor/rules/avoid-ai-writing.mdc
https://raw.githubusercontent.com/conorbronsdon/avoid-ai-writing/main/cursor-rules/avoid-ai-writing.mdc
- Hermes: place SKILL.md in Hermes skills directory with a curl example:
mkdir -p ~/.hermes/skills/writing/avoid-ai-writing
curl -o ~/.hermes/skills/writing/avoid-ai-writing/SKILL.md
https://raw.githubusercontent.com/conorbronsdon/avoid-ai-writing/main/SKILL.md
- OpenAI Codex: place SKILL.md under .agents/skills/avoid-ai-writing or ~/.agents/skills/ with curl:
mkdir -p .agents/skills/avoid-ai-writing
curl -o .agents/skills/avoid-ai-writing/SKILL.md
https://raw.githubusercontent.com/conorbronsdon/avoid-ai-writing/main/SKILL.md
Triggered usage phrases include:
- "Remove AI-isms from this post"
- "Audit this draft for AI tells"
- "Make this sound less like AI"
- "Clean up AI writing in this paragraph"
In rewrite mode, the tool returns four sections: Issues found, Rewritten version, What changed, and Second-pass audit. In detect mode, it returns Issues found and Assessment.
## Recent releases
Latest releases (5 most recent):
- v3.23.0 v3.23.0 (2026-08-04): Added an optional `--style` house-style layer. The note mentions applying a user-supplied config (`register` directives the model follows, plus `mechanics`).
- v3.22.3 (2026-08-03): Changed five prose-contract clarifications in SKILL.md; no rule or detector behavior changes.
- v3.22.2 (2026-08-03): Fixed: `hashtag-stuff` counted every `#word`, correcting false positives.
- v3.22.1 (2026-07-31): Fixed: `title-case-header` firing on Markdown heading; adjusted anchor logic.
- v3.22.0 (2026-07-31): Added two pieces of enforcement; catalog remains at 61 with engine type count increase from 46 to 47.
## Traction
Stars: 2794. Forks: 257. Open issues: 23. Language: JavaScript. License: MIT. Created: 2026-03-06. Last push: 2026-08-04.
## Behind the repo
This section would include linked startup/company details if present in the FACTS, but none are provided beyond repository and tool scope.
## Caveats
License: MIT. Created: 2026-03-06. Last push: 2026-08-04. Open issues: 23. Language: JavaScript. No additional age or security notes are provided in the facts.