autonomyx-fraud-sentinel

Fraud detection, prevention, reporting, and public awareness for Autonomyx. Collects signals from agents and users and detects seven patterns: scarcity manipulation, dormant-identity resurrection, velocity spikes, new-device anomalies, bad-actor networks, time-of-day anomalies, and low-value probe transactions. SurrealDB active engine — DEFINE EVENTs fire on every write and LIVE SELECT pushes alerts in real time. Enrichment via FRI (India), IPQS, SEON, SpyCloud, and NPCI MuleHunter.AI. Flags IDENTITY RISK rather than amount — even ₹1 to a suspicious identity is flagged. Pre-action scoring returns ALLOW / WARN / STRONG_WARN / BLOCK in under 200 ms. Citizen Safety Check: describe a transaction to receive a TXN ID — zero PII stored; the token can be used as evidence for bank claims, cybercrime.gov.in, or a police FIR. Ingestion and alerting ALWAYS trigger when any of these appear: "fraud", "scam", "suspicious", "fake", "impersonation", "dormant account", "sudden urgency", "probe transaction", "small payment", "access block"; when a user reports fraud; when an agent submits a signal; or when a new platform agent is being onboarded.