Zero-Trust Traffic Governance for Kubernetes AI Agent Sandboxes
Talkat KubeCon + CloudNativeCon + OpenInfra Summit + PyTorch Conference China 2026 · 2026-09-07..09-09·
summary
Describes a zero‑trust network control path built for Kubernetes AI agent sandboxes that enforces per‑request destination, identity, credential boundaries, and auditing. Explains concrete mechanisms: domain/service/network target restrictions, combining platform guardrails with tenant allowlists, handling DNS timing and short‑lived sandboxes, and doing identity injection, token replacement, LLM request auditing, and forced forwarding via an internal model gateway without a heavy per‑sandbox sidecar.
AI-written summary · about
related talks
42mZero Trust for Autonomous Agents: Isolating AI Workloads on Kubernetes
29mZero Trust for APIs: From Edge to Mesh with Istio
45mZero Trusting as a True Cloud Native Dev
29mZero Trust Security with Service Mesh
32mImplementing Zero Trust in a Kubernetes Cluster with Istio
39mTrust, but Sandbox: Securing AI Agents in Kubernetes
01h 02mIntroduction to Zero Trust Security
35mDemystifying Zero-Trust for Cloud Native Technologies