I parsed 559 public CLAUDE.md files. Most of what's in them isn't rules.

I've been building a tool that checks whether Claude Code actually followed the rules in your CLAUDE.md. Early on I realised I was testing it against one file: my own. So I went and got 559 real ones.

CLAUDE.md, AGENTS.md, .cursorrules, Copilot instructions, Windsurf, Gemini — from PyTorch, Kubernetes, Elasticsearch and a few hundred smaller repos. 21,986 items parsed out of them, no crashes.

Two things came out of it that I didn't expect.

63.4% of a rules file isn't a rule

Directory listings. Reference tables. Import examples. Architecture notes. Glob syntax cheatsheets.

I parsed 559 public CLAUDE.md files. Most of what's in them isn't rules.

Related reading

What actually belongs in CLAUDE.md — and what to move to skills, hooks, or docs

Claude Code’s Secret Weapon: A Complete Guide to CLAUDE.md | Towards AI

How to Keep Your CLAUDE.md Light (and Why More Rules Make Agents Worse)

My Claude Code setup in 2026

CLAUDE.md for an iOS Team: What to Put In It (and What to Leave Out)

5 ways your CLAUDE.md rules quietly fail