This week: 16% of 499 scanned repos had CLAUDE.md.See Radar #4 →

What it is

CLAUDE.md is a Markdown file at the project root (or in subfolders) that Claude Code loads as persistent context — describing architecture, commands, and rules for the assistant.

Why use it

It improves answer quality and lets you encode hygiene conventions, such as instructing the assistant to avoid reading or echoing secret files. It pairs well with .claudeignore for actual exclusion.

How to set it up

Create a CLAUDE.md in your project root and document setup commands, conventions, and any files the assistant should treat as off-limits.

Example CLAUDE.md
# Project notes

## Hygiene
Do not read or print .env, *.key, or *.pem files.

What to exclude

Common sensitive categories worth adding to CLAUDE.md:

FAQ

Does CLAUDE.md block access to files?
No. It provides instructions, not enforcement. Use .claudeignore to actually exclude sensitive files from context.

Check your own project

158 of 499 repos this week exposed sensitive paths to AI tools. Check yours in seconds:

Run the same local check with Offsend CLI:

brew install --cask offsend/tap/offsend-cli
offsend show
offsend prepare --dry-run

Everything runs locally. No file contents are uploaded.