As autonomous AI agents built with Microsoft Foundry take on critical tasks, security is non-negotiable. In this fast demo, we apply zero trust to agent chains: secure tool access with Entra ID, protect secrets and context with Key Vault, enforce real-time content filters, and capture auditable traces. Learn to stop prompt injection, data leaks, and agent hijacking using Azure’s native security controls. Walk away with a battle-tested checklist for hardening agents in Foundry!