Run /init to Generate a CLAUDE.md Tailored to Your Project Automatically
Every project benefits from a CLAUDE.md, but writing one from scratch feels like a chore. Let Claude do it instead — /init analyzes your actual project and generates one that's already useful.
> /init
Claude scans your codebase and produces a CLAUDE.md that includes:
- Build and test commands — detected from your package.json, composer.json, Makefile, or CI config
- Project architecture — directory structure, key patterns, and how the pieces fit together
- Coding conventions — style rules inferred from existing code (naming, formatting, patterns in use)
- Tech stack — frameworks, libraries, and tools it detected
The generated file lands in your project root, ready to commit. From that point on, every Claude Code session in this project starts with that context — Claude knows your commands, respects your conventions, and understands your architecture without you explaining it each time.
After generation, review and refine it. The auto-generated version is a solid starting point, but you'll want to add things Claude can't infer:
# Things to add manually after /init
- Team-specific conventions ("we always use form requests, never validate in controllers")
- Deployment notes ("staging deploys on push to develop branch")
- Known gotchas ("the legacy billing module uses a different DB connection")
- Preferred approaches ("use Pest for tests, not PHPUnit directly")
Commit the CLAUDE.md to version control so your entire team benefits. When anyone on the team opens Claude Code in this repo, they get the same baseline context.
Five seconds of /init saves five minutes of explanation at the start of every session.
via Claude Code
Log in to leave a comment.
When Claude writes error messages, button labels, validation text, or onboarding flows, it defaults to generic developer-speak. Add a "Users" section to your CLAUDE.md describing who your actual users are — their technical level, industry jargon, and what they care about — so Claude writes copy that makes sense to THEM, not to developers.
Use the --agent flag with custom markdown files in .claude/agents/ to launch purpose-built Claude sessions with restricted tools and scoped system prompts.
Every project has traps — the billing module that silently fails if you forget to queue the job, the legacy table with column names that don't match the model, the config value that must be set before tests run. Document these gotchas in your CLAUDE.md so Claude avoids the same mistakes your team spent days debugging.