{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/457fc49585834d2fa1eb8cf846a38e6e\" frameborder=\"0\" width=\"1920\" height=\"1440\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":1440,"width":1920,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":1440,"thumbnail_width":1920,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/457fc49585834d2fa1eb8cf846a38e6e-15a1aedcc763a986.gif","duration":286.252633,"title":"Zenity Agent Security with AWS Cedar Guardrails","description":"This Loom explains how Zenity secures AWS AI agents using runtime guardrails built on AWS Cedar. It covers continuous inventory, risk assessment, and runtime monitoring across Bedrock agents, Agent Core, and custom agents, reconstructing full agent trajectories when threats occur. Zenity uses formally verifiable policies to allow, modify, escalate, or block actions based on acting and affected identities, detecting issues like prompt injection, jailbreaks, and sensitive data exfiltration. The speaker cites blocking 14 unauthorized escalation attempts across more than 4200 evaluated requests with checks under 30 milliseconds."}