Can I vibe code Eye4Fraud?

Build a internal E-commerce Fraud Screening Microservice in Node.js/TypeScript using Express and PostgreSQL.

1. DATA MODEL:
- Orders: id, platform_order_id, total_amount, currency, customer_email, billing_address, shipping_address, customer_ip, status (APPROVED, REJECTED, MANUAL_REVIEW), risk_score (0-100).
- FraudRules: id, rule_name, field, condition, threshold_value, risk_weight, active.
- AuditLogs: id, order_id, matched_rule, score_delta, timestamp.

2. CORE FUNCTIONALITY:
- Expose POST /api/v1/screen-order webhook that accepts standard Shopify order payloads.
- Implement a heuristics scoring engine that checks:
  a) AVS/CVV result codes (penalize mismatches).
  b) Shipping address vs billing address match.
  c) Velocity checks: Flag >2 orders from same IP or email in 60 minutes.
  d) High value check: Orders > $500 add baseline risk score.
- Integration with MaxMind minFraud API (optional mock interface for dev) to fetch IP risk score.
- If cumulative risk score >= 75, mark status MANUAL_REVIEW and trigger a Slack webhook alert.
- Expose GET /api/v1/review-queue for internal ops to approve or reject held orders.

3. FAILURE MODES TO HANDLE:
- If MaxMind API times out, fallback to local rule-based evaluation and log an API warning.
- Idempotency: Ignore duplicate webhooks for the same order_id using Redis locking.
- Handle missing or malformed address fields without throwing uncaught exceptions.

4. OUT OF SCOPE:
- Chargeback financial reimbursement or liability insurance.
- Machine learning model training pipeline.
- Direct payment gateway authorization/capture manipulation.

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