Bootstrapping a Project with /init

The first thing to do in a new project is teach Claude about it. The /init command does this for you by creating a CLAUDE.md file — a set of standing instructions Claude loads every time it starts in that folder.

What it does

Run it once in your project root:

/init

Claude looks over your codebase and writes a starter CLAUDE.md describing what the project is, how it's built, and how to run it. Instead of re-explaining your stack in every session, you write it down once and Claude reads it automatically at launch.

Why it matters

CLAUDE.md is loaded at the start of every session, so anything in it shapes how Claude behaves without you repeating yourself. Good things to capture:

• What the project is and its main pieces.
• The commands to build, test, and run.
• Conventions and patterns Claude should follow.

Keep it tight

The generated file is a starting point, not the final word. Read it and trim it. A few tips that pay off:

• Keep it short — aim for under about 200 lines.
• Prefer concrete rules over vague guidance.
• Don't document what Claude can already infer from the code itself.

Once your CLAUDE.md exists, you can browse and edit it any time with /memory, or drop a quick note into it by typing # followed by your reminder. Treat it as a living file: when you notice Claude repeatedly missing something about your project, add a line so it stops missing it. A well-kept CLAUDE.md is the single biggest lever on how well Claude fits your codebase.