NVIDIA launches OpenShell, an open secure runtime that governs AI agent execution outside the model
TheZachMueller · x · 2026-09-30
NVIDIA unveiled OpenShell, an open secure runtime for AI agents and the centerpiece of its Open Agent Safety Platform.
Key points:
- Security lives in the environment, not the model or app: deny-by-default, policy-based permissions enforced outside the agent's process, so controls can't be prompted away; every allow/deny is auditable.
- Model- and harness-agnostic: any model, any agent framework, governed through one unified policy layer.
- Flexible deployment: cloud, hybrid, on-premises, edge, and air-gapped environments.
An accompanying blog covers adding runtime controls to existing agents without rewriting their harness—protecting credentials, restricting access, and reviewing policy changes. The author plans to try it on a few idle Raspberry Pis for lightweight workloads.
More from coding & agent
- AI Agent Sets 7 Records in OpenAI Hiring Challenge, Doubling Best Human's 3 — victor_explore · 2026-09-30
- Dev Launches Anonymous Tech Confessions Site, AI Coding Sins Steal the Show — ayushtweetshere · 2026-09-30
- Codegraph rockets up GitHub Trending with 72k stars: pre-indexed code knowledge graph for AI coding agents — colbymchenry · 2026-09-30
- MCP official servers repo nears 91k stars on GitHub as ecosystem keeps growing — modelcontextprotocol · 2026-09-30
- Ex-Googler: TensorFlow is in maintenance mode — here's how to migrate to PyTorch — lmoroney · 2026-09-30
- Adding a shared memory layer cut this multi-agent bill from $580 to $260 a month — montemom · 2026-09-30