Google Cloud Launches "Malicious Skill" AI Threat Detectors for Kubernetes Workloads

Google Cloud has rolled out new runtime defenses, dubbed "Malicious Skill" threat detectors, for artificial intelligence workloads running on Google Kubernetes Engine (GKE), Cloud Run, and its Agent Platform. These detectors monitor the actual behavior of AI agent code in production, rather than just scanning container images, to catch malicious activity post-deployment.