NVIDIA Releases an Open-Source Agent Security Platform with Chip-Level Continuous Monitoring and Isolation
NVIDIA Open Agent Safety Platform: A Reference for Continuous In-Silicon Agent Monitoring
AI BriefingNVIDIA introduced the NVIDIA Open Agent Safety Platform, comprising the open-source OpenShell runtime, the hardware-layer Sentry, and DOCA technology. The platform is designed to prevent AI agent behavior drift and privilege escalation through kernel-level isolation, zero-trust environments, and out-of-band monitoring. OpenShell converts operational instructions into verifiable policies, while the BlueField DPU provides real-time policy enforcement and identity governance along the model path, ensuring that the system can be protected independently even when the host is untrusted.
11 additional sources reported this
- Anthropic and NVIDIA Partner to Launch Claude Managed Agents and OpenShell, Strengthening Agent Security ControlsClaude: Blog · September 29
- NVIDIA Releases Open Agent Safety Platform: OpenShell Sandboxes Agents, While Sentry Uses BlueField-4 for Millisecond-Level IsolationMarkTechPost · September 29
- Nvidia Releases Open Agent Safety Platform, Using Independent Software and Hardware Security Layers to Constrain Rogue AI AgentsTechCrunch: AI · September 29
- Hugging Face CEO Shares a Postmortem of an Agent Network Attack and Open-Sources OpenShell Egress Traffic MonitoringClément Delangue · September 28
- Nvidia Releases Sentry Hardware Watchdog to Isolate Rogue AI Agents Using BlueField-4The Decoder: AI News · September 28
- Nvidia Launches Open Agent Safety Platform, Able to Isolate Out-of-Bounds AI Agents in MillisecondsThe Verge: AI · September 28
- Nvidia Releases Open-Source AI Agent Safety Platform, Featuring OpenShell Sandboxing and Sentry Hardware IsolationRohan Paul · September 28
- Nvidia Releases AI Agent Safety Platform Featuring Sentry for Real-Time Isolation of Anomalous AgentsITHome · Artificial Intelligence · September 28
- NVIDIA Releases Open-Source Runtime OpenShell, Providing AI Agents with Permission Controls and a Secure SandboxNVIDIA Technical Blog: Agentic AI / Generative AI · September 28
- NVIDIA and More Than 100 Partners Launch Open Agent Safety PlatformJensen Huang · September 28
- NVIDIA Releases Open Agent Safety Platform to Protect Agents with OpenShell and SentryNVIDIA: Newsroom · September 28
Source: NVIDIA Technical Blog: Agentic AI / Generative AI · developer.nvidia.com
Event follow-up1
View all event coverage- NVIDIA Releases an Open-Source Agent Security Platform with Chip-Level Continuous Monitoring and IsolationNVIDIA Technical Blog: Agentic AI / Generative AI · September 28