The Avoid AI Writing skill
Strip AI tells out of any draft. Setup for claude.ai and Claude Code.
From the reel: SKILL · Updated Oct 2026
What it does
Audits and rewrites text to remove the tells that make content read as AI-written: promotional filler, hedge-stacked predictions, "serves as" and "boasts", chatbot artifacts ("Certainly!", "I hope this helps!"), and over-uniform sentence rhythm. It checks 42 pattern categories.
Two modes:
- Rewrite (default): flags issues, rewrites, then runs a second pass.
- Detect: flags patterns without changing anything.
GitHub: github.com/conorbronsdon/avoid-ai-writing
Option 1: Upload to claude.ai or Claude Desktop
No command line needed.
- Open the GitHub repo.
- Click the green Code button → Download ZIP.
- Unzip it. GitHub names the folder
avoid-ai-writing-main. Rename it toavoid-ai-writing, then zip that folder so SKILL.md sits at the root. Getting this right avoids the most common upload error. - In Claude, go to Customize → Skills, click "+", then "Create skill".
- Upload the ZIP and toggle the skill on.
Requirements: custom skills need "Code execution and file creation" turned on, and are available on Pro, Max, Team and Enterprise plans. Skills you upload are private to your account.
Option 2: Clone with Git (Claude Code)
Clone straight into your Claude Code skills folder:
git clone https://github.com/conorbronsdon/avoid-ai-writing ~/.claude/skills/avoid-ai-writing
Claude Code picks it up automatically.
Or clone anywhere and reference it in your CLAUDE.md:
git clone https://github.com/conorbronsdon/avoid-ai-writing
- Editing for AI patterns → read `path/to/avoid-ai-writing/SKILL.md`
Using it
Just ask:
- Remove AI-isms from this post
- Audit this draft for AI tells
- Make this sound less like AI
Add "detect", "flag only" or "scan" to get flags without a rewrite.
Notes
- Skills uploaded to claude.ai don't auto-update from GitHub. Re-download and re-upload to get new versions.
- MIT licensed. Made by Conor Bronsdon.