Running untrusted agents on GKE: sandboxing, cold starts, and state
rseroter · x · 2026-11-19
Traditional unit-tested, scanned code in Kubernetes is fine — but agents write and execute code on the fly, breaking the trust model. A diagram breaks down three risks and fixes: (1) unreviewed LLM-generated code on a shared host kernel risks container escape and credential theft → GKE Sandbox (gVisor) or MicroVM (Kata+KVM); (2) multi-second pod spin-up breaks interactive tool loops → pre-warmed SandboxWarmPool with <1s SandboxClaim; (3) idle agents waste RAM but deletion wipes state → Agent Substrate RAM snapshot suspend/restore in <500ms. Includes a live demo of untrusted agents on Agent Substrate with GKE.
More from coding & agent
- Google Labs launches Stitch Loop, an always-on product agent for post-launch improvement — GoogleLabs · 2026-11-19
- Same models, more bugs: Xint's harness finds 5 critical nginx bugs, 3 never found before — dyn___ · 2026-11-19
- Neo4j Open-Sources Relate: Embedded Rust Graph DB for AI Agent Memory — JeremyCMorgan · 2026-11-19
- lineup: Open-Source CLI That Catches Coding Agents Making Sneaky Edits — Several-Coyote3534 · 2026-11-19
- Wanted: an LLM Assistant Inside ComfyUI — External MCP Feels Too Clunky — dustynjerman · 2026-11-19
- Agentrooms: An MCP Server Letting Claude Code Ask Codex to Review Commits — terzigolu · 2026-11-19