Context : Expand AI is fully deployed and planned; we’re integrating backends and customizations.
Stack: Next.js + LangChain + Supabase (+ LangGraph as needed) .
Access & Setup
Accept GitHub Invite
Check your email for the GitHub organization invite and accept it
Clone Repository
git clone https://github.com/Weventures-AI/expand-ai.git cd expand-ai
Read Documentation
Review these three critical files: cat README.md cat claude.md cat plan.md cat agents.md
Environment Setup
# Create .env.local from template cp .env.example .env.local # Add your Supabase keys # Test DB connectivity
Install & Run
Confirm the app boots at http://localhost:3000 Execution Loop Your daily workflow for implementing features:
1. Planning
2. Implementation
3. Review
4. Ship
Open Claude Code # Start Claude Code terminal claude
Load Context Paste relevant excerpts from plan.md and current milestone requirements Request Implementation Plan Based on plan.md Phase 2, generate the scaffold for the user authentication flow with Supabase RLS
Generate Code Ask CC to create:
Component structure
API routes
Database migrations
Test scaffolds
Run Locally Debug with AI Feed console errors/logs back to CC for fixes: Getting this error when submitting the form: [paste error]
Codex Review Ask Codex to review your changes: Review this diff for security issues and suggest improvements: [paste git diff]
Write Tests # Unit tests pnpm test # E2E tests pnpm test:e2e
Ensure CI Passes All checks must be green before PR Commit Frequently git add . git commit -m "feat: implement user auth with Supabase RLS"
Open PR Early git push -u origin feat/user-auth gh pr create
Monitor Deployment Check Sentry and logs after merge Example Milestones Your first sprint will likely include:
Confirm auth flow works end-to-end
Add missing Supabase RLS policies
Test multi-tenant data isolation
Document auth architecture
M2: First Agentic Workflow
Implement LangChain tool calling
Add retry logic and fallbacks
Create workflow monitoring
Write integration tests
Create Supabase Edge Function
Integrate with frontend components
Add streaming support
Implement rate limiting
Configure Sentry error tracking
Add OpenTelemetry traces
Create performance dashboards
Set up alerts
Strengthen input validation with Zod
Improve error envelopes
Add rate limiting
Security audit
Definition of Done Common Commands Development
Database
AI Development
# Start development server pnpm dev # Run tests pnpm test pnpm test:e2e # Lint and format pnpm lint pnpm format
Getting Help Pro tip : Keep claude.md, plan.md, and agents.md open in your editor while working. They’re your north star for implementation decisions.
