Use this checklist to track your progress in setting up and maintaining your claude.md file.
- Downloaded or copied
claude.mdto my project root - File is named exactly
claude.md(lowercase) - File is in the project root (same level as package.json/README.md)
- Described what my app does (1-2 sentences)
- Selected my frontend framework
- Selected my UI/styling approach
- Selected my state management
- Selected my package manager
- Listed Supabase features I'm using
- Listed my main database tables
- Added my actual dev commands
- Committed claude.md to git
- Did NOT commit any secrets or API keys
- Added a commit message like "Add claude.md for AI context"
🎉 Congratulations! You now have a working claude.md file!
The rest is optional and can be added as you build your project.
Do these when you have more time:
- Confirmed TypeScript usage (Yes/No)
- Added deployment platform
- Listed all major dependencies
- Documented first-time setup steps
- Listed all environment variables needed
- Explained how to run dev server
- Added how to run tests (if applicable)
- Documented file naming patterns
- Added code style preferences
- Documented Supabase query pattern
- Added error handling approach
- Described each main table
- Documented key columns
- Explained access patterns
- Documented RLS approach
- Added at least ONE critical rule AI should know
- Documented what NOT to do
- Added security considerations
Add these as your project grows:
- Documented directory structure
- Explained architectural decisions
- Described state management strategy
- Documented authentication flow
- Explained data fetching patterns
- Documented key business rules
- Explained important workflows
- Added domain-specific terminology
- Listed current bugs/limitations
- Documented workarounds
- Added technical debt items
- Documented color scheme
- Added spacing/typography rules
- Listed reusable components
- Added new database tables to Database section
- Documented new patterns in Code Conventions
- Updated Business Logic if rules changed
- Added any gotchas to Known Issues
- Skimmed through entire claude.md
- Removed outdated information
- Updated technology stack if changed
- Added anything discovered this month
- Verified examples still work
- Full read-through of claude.md
- Updated all sections
- Removed completed TODOs
- Reorganized if needed
- Team review session held (if applicable)
- Pointed them to claude.md
- Asked them to read it
- Got feedback on what's unclear
- Updated based on their questions
Add these only when needed:
- Documented payment provider (Stripe, etc.)
- Added email service (SendGrid, etc.)
- Documented error tracking (Sentry, etc.)
- Added analytics setup
- Listed other integrations
- Documented monitoring setup
- Added logging approach
- Explained performance budgets
- Added analytics events
- Documented testing approach
- Explained how to run tests
- Added testing patterns
- Listed what we test/don't test
- Documented deployment process
- Added environment-specific configs
- Created deployment checklist
- Listed all environments
- Internationalization (if applicable)
- Feature flags (if using)
- Mobile considerations (if responsive/PWA)
- Security policies
- Design system
- Background jobs
- Webhooks
- Email templates
- Created
.env.examplewith all required variables - Updated README.md to mention claude.md
- Added "Update claude.md" to PR template (if applicable)
- Shared claude.md with team
- Got buy-in from team members
- Assigned section ownership (optional)
- Scheduled first review session
Run this check periodically to ensure quality:
- All sections have concrete examples (not just placeholders)
- No
[TODO]markers left unfilled - Examples are accurate and tested
- No outdated information
- Explains WHY not just WHAT
- Can be read in under 5 minutes (for Quick Start)
- Table of contents is accurate (if you added one)
- Sections are well-organized
- No duplicate information
- Links work (if any)
- No API keys or secrets
- No database passwords
- No internal URLs (if sensitive)
-
.env.examplehas placeholders, not real values
- Actually helps AI make better suggestions
- New team members find it useful
- Reduces repeated questions
- Matches current state of project
- Removed sections that don't apply to my project
- Deleted tech stack options I'm not using
- Removed advanced sections I don't need yet
- Added project-specific sections
- Included domain-specific context
- Added industry-specific patterns
- Adjusted tone to match team culture
- Added company-specific conventions
- Included team preferences
- Added to PR review checklist
- Made part of onboarding
- Scheduled regular reviews
- Team knows to update it
- Tested with Claude
- Tested with other AI tools (if using)
- Verified AI reads and uses it
- Confirmed improved AI suggestions
Track these to measure effectiveness:
- AI makes fewer mistakes
- AI asks fewer clarifying questions
- Generated code matches our patterns
- New team members reference it
- Team finds it valuable
- Count: PRs that update claude.md
- Track: Time to onboard new developer
- Measure: AI assistance effectiveness (subjective 1-10)
- Monitor: Questions in Slack/Discord that could be in claude.md
- Read README.md
- Read FILL_OUT_GUIDE.md
- Reviewed example.claude.md
- Skimmed QUICK_REFERENCE.md
- Browsed FAQ.md
- Checked ADVANCED.md (for future reference)
- Understand what makes good AI context
- Know when to update claude.md
- Can write good examples
- Can explain value to team
- Can maintain it over time
- Starred the repo (if useful)
- Shared with others
- Submitted bug reports/improvements
- Shared my experience
- Contributed to discussions
- Know where to ask questions (FAQ, Discussions, Issues)
- Subscribed to updates
- Joined community (if available)
Add your own project-specific items here:
- [Your checklist item]
- [Your checklist item]
- [Your checklist item]
- [Your checklist item]
- [Your checklist item]
- [Your checklist item]
✅ Quick Start section filled out ✅ File committed to git ✅ No secrets included
✅ Above + ✅ Code conventions documented ✅ Database tables listed ✅ Important rules added
✅ Above + ✅ Business logic explained ✅ Known issues documented ✅ Architecture decisions recorded
✅ Above + ✅ Kept up to date ✅ Team actively uses it ✅ Measurably improves AI assistance
Where are you at?
- Minimum Viable (good enough to start!)
- Good (solid foundation)
- Great (comprehensive)
- Excellent (maintained and valuable)
Next action: ___________________________
Review date: ___________________________
Print this checklist or keep it in your project for tracking progress!