StrongDM Delivers Policy Enforcement for Agentic AI with Leash

StrongDM released Leash, an open-source project that provides real-time policy enforcement for AI agents. Operating at the kernel level, Leash intercepts agent activity and evaluates it against Cedar-defined policies, offering visibility and control over autonomous workloads. The system monitors agent actions like file access and network calls, enforces context-aware rules before connections are established, and logs all activity for audit purposes. Leash extends StrongDM's access control model from human operators to machine identities, addressing the growing need to govern AI agents that execute unpredictable actions across enterprise networks.