Arbor

A certified operational data repository. Manufacturers, suppliers and producers upload operational documents; Arbor extracts the figures, has a person confirm them, and stores each one with its source, a trust tier (Verified, Declared or Estimated) and an HMAC-chained audit entry. Buyers request, query and share that data. CBAM cases are opened from confirmed customs documents and handled by the Nucleos engine in nucleos/.

Layout

Path	What it is
src/	The Next.js app (App Router), TypeScript, Prisma, Postgres
prisma/	Schema and migrations
nucleos/	The CBAM engine (FastAPI, Python) and its legacy front ends
brain/	Stateless maths service (calibration, fusion, privacy) — Python
contract/	The Arbor–Nucleos contract; TypeScript is generated from it
docs/	Deployment, integration rollout, security (WISP, risk register, incident response)

Running it

npm ci
cp .env.example .env        # fill in the values
npm run dev

Checking it

npx tsc --noEmit && npx eslint src && npx jest     # Arbor
npm run python:setup                                # once: creates nucleos/.venv
npm run test:python                                 # Nucleos (set TEST_DATABASE_URL for Postgres — see nucleos/scripts/create_test_db.sh)
npm run contract:check && npm run verify:boundary   # Arbor–Nucleos contract and HTTP boundary

CI runs all of these on every pull request.

Deploying

Migrations are applied as a separate, deliberate step before the deploy that needs them — never by the build. See docs/DEPLOYMENT.md. The Nucleos rollout and its open items are in docs/INTEGRATION-ROLLOUT.md.

Rules

CLAUDE.md holds the architecture rules (three layers, trust tiers, design system) and the deployment gate.