Can I vibe code FraudLabs Pro?

Build a lightweight Node.js/TypeScript microservice using Express and Prisma to screen e-commerce orders for fraud.

1. DATA MODEL:
- Store Orders: order_id, customer_email, total_amount, currency, shipping_country, billing_country, ip_address, risk_score, status (APPROVED, REVIEW, REJECTED), flags (jsonb).
- Store Rules: id, name, condition_field, operator (EQUALS, NOT_EQUALS, GREATER_THAN, IN_LIST), threshold_value, risk_points, active (boolean).
- Blacklist: type (EMAIL, IP, DOMAIN, COUNTRY), value, reason.

2. CORE FUNCTIONALITY:
- Expose POST /api/v1/screen-order endpoint accepting order JSON payloads.
- Implement a rule evaluation engine:
  a. Check payload against Blacklist table. If match found, set risk score directly to 100.
  b. Evaluate location: Compare IP geolocated country (using a local GeoLite2 database or mock service) against billing_country and shipping_country. Add 30 points if mismatched.
  c. Check disposable email domain list (parse email domain against a configured list). Add 25 points if match found.
  d. Execute custom dynamic rules from the Store Rules table and aggregate risk points.
- Determine status:
  - Score < 30: APPROVED
  - Score 30-69: REVIEW
  - Score >= 70: REJECTED
- Expose POST /api/v1/orders/:id/approve and /reject endpoints to allow human overrides.

3. FAILURE HANDLING & TIMEOUTS:
- Wrap evaluation execution in a 500ms timeout. If external calls fail or time out, fail open by marking status as REVIEW with a system error flag rather than blocking the checkout flow.
- Ensure email parsing handles edge cases (e.g., malformed email strings, subdomains).

4. OUT OF SCOPE:
- Machine learning model training.
- Direct payment gateway API chargeback dispute integrations.

$ each button prefixes agent-specific run instructions · build your own product, never copy proprietary code, trademarks or designs