PRE Security Launches AgentGuard for AI Agent Monitoring

PRE Security launches AgentGuard, a security capability that monitors autonomous AI agents and blocks risky actions, the company said in a press release. The product is part of the PRE Security Platform and treats each agent as an identity with a defined mission, behavioral history, and expected activity.

AgentGuard runs between an AI agent and the tools it accesses. Its AI Agent Behavioral Analytics system evaluates the agent's identity, mission, resources, tools, and sequence of actions. It also uses PRE Threat Intelligence when assessing activity such as contact with malicious domains.

At execution, AgentGuard can allow an action, monitor it, request human approval, or block it. PRE Security said the system can identify a sequence of individually valid actions that collectively indicate activity such as data theft.