openclaw-fraud-detectlisted

# Openclaw Fraud Detect ## Overview Create a small FastAPI webhook service that accepts Stripe fraud-related events, verifies signatures, summarizes risk signals, and forwards the summary to OpenClaw for memory storage and alerting. ## Workflow ### 1. Confirm the baseline implementation Start from the existing reference implementation and keep changes minimal: - `/home/coder/fraud-detects/fraud_webhook/app.py` ### 2. Verify Stripe webhook signatures Use `stripe.Webhook.construct_event` with the `Stripe-Signature` header and `STRIPE_WEBHOOK_SECRET`. Reject invalid signatures with a 400 error. ### 3. Filter to fraud-signal events Use an allowlist to keep noise low. Default allowlist: - `radar.early_fraud_warning.created` - `charge.dispute.created` - `review.closed` Only process events on the allowlist and return a structured "ignored" response otherwise. ### 4. Summarize the event Produce a compact JSON summary that includes: - event id, type, created, livemode - charge / payment_intent / customer identifiers - amount and currency - risk fields (`risk_level`, `risk_score`, `seller_message`, `network_status`, `reason`) if present Load detailed field guidance from `references/stripe-fraud-summary.md` if needed. ### 5. Forward to OpenClaw If the gateway token is present, send: - `memory_store` via `/tools/invoke` for storage - a `/v1/responses` request instructing Slack alerting via Civic Nexus tools Use the environment variables and request shapes in `references/o