Session correlation
Session correlation is the analysis of an AI agent’s behavior across an entire session rather than one message at a time, detecting patterns no single message reveals: gradual boundary-pushing, intent that assembles across turns, retry storms after refusals, and drift from the agent’s original task.
Per-message screening judges each request in isolation, and sophisticated misuse is built to survive exactly that: each step innocuous, the sequence hostile. Correlation keeps state per session, scoring trajectories instead of snapshots, and can escalate a session (tighter review, then quarantine) when the pattern warrants. It is the stateful layer of the agent security stack.
Related
Shrike governs AI agent actions in real time: every command, query, and tool call evaluated against policy before it executes. Start with what is action governance or the quickstart.