CrowdStrike’s AgentWorks framework enables AI agents to operate autonomously across the Falcon platform and 1,800+ integrated third-party applications. The governance risk is structural: AI agents inherit service account or user-level permissions and operate in SOC infrastructure without the same access controls and detection coverage applied to human operators. No CVE, active exploitation, or KEV listing is associated with this item.