High-Priority 2026 AI Threat Detection — Non-Browser Outbound Connections to LLM API Endpoints
Detects outbound network requests to popular Large Language Model (LLM) API endpoints originating from processes that do not appear to be standard web browsers or identified legitimate development tools. This behavior may indicate an unauthorized script, malware, or compromised process attempting to use LLM APIs for C2, data exfiltration, or automated reconnaissance.
Sigma
