Nvidia launches the Open Agent Safety Platform, a reference design to stop AI agents from escaping, made up of OpenShell for CPUs and Sentry for network chips (@kifleswing / CNBC) (Visit Techmeme dot com for the link and full context!)

Introducing the NVIDIA Open Agent Safety Platform. An open reference design built with partners that continuously monitors and governs agent behavior, ensuring that AI agents follow the rules. Read the release:

Run the agent. Secure the infrastructure around it. As AI agents become more autonomous, security has to be built into the infrastructure around them. Arm CPUs run agentic workloads, while Arm-based DPUs provide independent observation, isolation and policy enforcement. @NVIDIA’s Open Agent Safety Platform shows that model in action. 🔐

$NVDA LAUNCHES OPEN AGENT SAFETY PLATFORM Nvidia is introducing OpenShell software and the Sentry reference system design to let companies set enforceable limits on what AI agents can access and do, with security controls operating outside the agents themselves. OpenShell, the open-source software layer, controls access to files, tools and external services, protects credentials and records permission decisions. It works with existing agents such as Claude Code, Codex and Hermes without requiring developers to rewrite them. Sentry adds a separate hardware watchdog running on Nvidia’s BlueField-4 processors. Nvidia says it continuously monitors agent activity from outside the agent’s environment and can quarantine agents attempting to cross their boundaries within milliseconds. The approach is designed for agents that work for hours or days, write code and interact with business systems. Instead of relying only on the model to follow instructions, the infrastructure limits which actions it can actually carry out. Anthropic has collaborated with Nvidia to add protections to Claude Managed Agents, while SpaceXAI is using the platform for Cursor coding agents and Grok models. Salesforce has integrated OpenShell with Slack, and SAP is embedding it into Joule Studio. Nvidia says more than 100 organizations are working with the technologies, including Microsoft, CrowdStrike, Cisco, Palo Alto Networks, Palantir and ServiceNow. The effort also extends to robotics, banks and energy infrastructure. Jensen Huang: “Safety and security require full-stack engineering.” OpenShell is broadly available, with platform software accessible through Nvidia’s developer resources and GitHub. Sentry is presented as a BlueField-4-based reference system design.

$ARM just showed $NVDA Open Agent Safety in action with “Run the agent. Secure the infrastructure around it.” OpenShell runs on CPU while Sentry on BlueField can shut down a bad agent giving Nvidia a reason to make DPUs standard in agentic AI racks.

Announcing the integration of the @NVIDIA Open Agent Safety Platform into the Dell AI Factory with NVIDIA. Infrastructure-level governance helps customers establish enforceable boundaries for autonomous agents, enabling secure, scalable agentic AI deployments with greater confidence and control.

We’ve launched NVIDIA Open Agent Safety Platform to help people control what AI agents can access and do. Agents can write code, use tools, and work on complex tasks for hours or days. That work requires access to data and systems, along with clear limits on how they’re used. NVIDIA OpenShell enforces permissions around the agent’s work. BlueField-4 and DOCA add independent monitoring and security controls in the infrastructure, outside the agent’s reach. Vera CPUs power the work itself. Together, these technologies give teams a foundation for putting agents to work with defined permissions, oversight, and protection. Explore the platform: