A concrete example
In CLAUDE.md I note the project's stack, the tone of voice, and the useful commands: Claude reads it first every time it opens.
Why it matters
Without CLAUDE.md, Claude rediscovers the project every session. With it, it follows your rules from the very first reply.
You see it at the root of a project, inside .claude folders, in agent tutorials, and in shared configs.
Don't mix it up with
Claude Code: Claude Code is Claude's agent: it reads a project, edits files, and runs commands with your approval.
Skill & Agent: A skill is a reusable method. An agent is a system that pushes a mission forward with tools, context, and decisions.
Common mistakes
- Writing a novel instead of actionable instructions.
- Putting secrets (API keys, passwords) in it instead of a .env file.
- Never updating it as the project evolves.
Quick checklist
- First I check whether the word names a concept, a tool, a risk, or a metric.
- I tie it to a concrete case: In CLAUDE.md I note the project's stack, the tone of voice, and the useful commands: Claude reads it first every time it opens.
- I keep the main trap in mind: Writing a novel instead of actionable instructions.
Quick questions
What is CLAUDE.md in AI?
CLAUDE.md is the file where you write down a project's rules so Claude doesn't start from scratch.
Where will I run into CLAUDE.md?
You see it at the root of a project, inside .claude folders, in agent tutorials, and in shared configs.
Which word should I read next?
Start with Claude Code, Skill & Agent, Slash command.