Who’s Really Shopping? Retail Fraud in the Age of Agentic AI

Unit 42 collection (March 2026) highlighting emergent AI-enabled threat trends and active exploitation: multiple research posts show widespread fragility of LLM guardrails (prompt fuzzing, AI-judge bypass, web-based indirect prompt injection) and growing weaponization of AI in malware and retail fraud. The feed also documents active nation‑state and criminal activity — Iranian-linked groups (Boggy Serpens, Handala Hack/Void Manticore) increasing wiper attacks and social-engineering misuse of Intune, a long-running espionage cluster (CL-UNK-1068), and suspected China-linked operations targeting