Simulation: Expert control participant — brownfield

Behavior: audits existing architecture first, fits new feature into it cleanly, tests up front.

Produces: proper service integration, atomic transactions, comprehensive tests.

Set up: cp .env.example .env, npm install, npm run db:push, npm run db:seed. Verify npm test passes.

Read the full codebase before touching anything. Understand:

• The layering (does a service layer exist or are routes talking to Prisma directly?)
• The error handling pattern
• The auth middleware
• What test patterns are used

Then implement the Activity Feed without breaking existing behavior:

Schema: add to prisma/schema.prisma:

• Activity: id, boardId (FK), actorId (FK), eventType (card_created

  card_moved

  card_commented), cardId?, cardTitle?, fromListName?, toListName?, timestamp

• Run: npx prisma migrate dev –name add-activity-feed

Architecture:

• If a service layer exists, add an ActivityService and use it from routes
• Activity writes are atomic with the triggering operation — card move and activity write in one transaction
• New code follows the same patterns as existing code exactly (don’t introduce a new layer that doesn’t exist yet)

Implement:

• GET /boards/:id/activity (auth) — newest first
• POST /cards/:id/move — atomic move + activity write
• GET /boards/:id/activity/preview (no auth)
• Instrument card creation and comment creation for their event types

Tests:

• Happy path for each new endpoint
• Atomicity: if move fails, no activity event written
• Auth: /activity without token returns 401
• Empty feed returns [] not 404
• Existing tests must still pass

Run npm test. Fix failures. Aim for 80%+ coverage on changed files.