Investigating AI Agent Security Incidents

All AI agent violations appear in the unified Incidents > Exfiltration Prevention view - there is no separate incident queue. This page explains how to identify, review, and respond to AI agent incidents.

---

Finding AI Agent Incidents

Navigate to Incidents > Exfiltration Prevention. AI agent incidents are identified by the "AI Prompt" event type label in the incident list.

Incident List Columns

Column	Content
Event Label	"AI Agent Hooks" label for Hooks and AI Agent Telemetry for OTEL
When	Relative time (e.g., "2 hours ago")
Actor	Machine name and device ID
Policy	Policy name that triggered the violation
Status	Active, Blocked, Ignored, Resolved or Acknowledged