# Open Agent Safety Platform

Published articles for Open Agent Safety Platform.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Inside NVIDIA Sentry: Secure AI Agents at Runtime

DevFeed: [Inside NVIDIA Sentry: Secure AI Agents at Runtime](<https://devfeed.tech/articles/inside-nvidia-sentry-secure-ai-agents-at-runtime-92595.md>)

Original publisher: [Read original article](<https://developer.nvidia.com/blog/inside-nvidia-sentry-secure-ai-agents-at-runtime/>)

Author: Ofir Arkin

Published: 2026-10-21T09:52:16Z

Content type: article

Language: en

Sources: [NVIDIA Developer](<https://devfeed.tech/sources/nvidia-developer.md>), [NVIDIA Technical Blog](<https://devfeed.tech/sources/nvidia-technical-blog.md>)

Topics: [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [DOCA](<https://devfeed.tech/topics/doca.md>), [OpenShell](<https://devfeed.tech/topics/openshell.md>), [Nvidia](<https://devfeed.tech/topics/nvidia.md>)

Tags: [a2a](<https://devfeed.tech/tags/a2a.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agentic-ai-generative-ai](<https://devfeed.tech/tags/agentic-ai-generative-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-networking](<https://devfeed.tech/tags/ai-networking.md>), [can](<https://devfeed.tech/tags/can.md>), [complete](<https://devfeed.tech/tags/complete.md>), [cross](<https://devfeed.tech/tags/cross.md>), [gtc-berlin-2026](<https://devfeed.tech/tags/gtc-berlin-2026.md>), [inside](<https://devfeed.tech/tags/inside.md>), [networking-communications](<https://devfeed.tech/tags/networking-communications.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-sentry](<https://devfeed.tech/tags/nvidia-sentry.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [runtime](<https://devfeed.tech/tags/runtime.md>), [secure-ai](<https://devfeed.tech/tags/secure-ai.md>), [secure-ai-agents](<https://devfeed.tech/tags/secure-ai-agents.md>), [security](<https://devfeed.tech/tags/security.md>), [security-controls](<https://devfeed.tech/tags/security-controls.md>), [sentry](<https://devfeed.tech/tags/sentry.md>), [show](<https://devfeed.tech/tags/show.md>), [top-stories](<https://devfeed.tech/tags/top-stories.md>), [trustworthy-ai-cybersecurity](<https://devfeed.tech/tags/trustworthy-ai-cybersecurity.md>), [using](<https://devfeed.tech/tags/using.md>)

### AI overview

NVIDIA Sentry places agent monitoring and policy enforcement on BlueField-4, outside the host and agent's control, to protect against agents that bypass software boundaries. DOCA Flow applies silicon-accelerated layer-4 network filtering; permitted traffic passes to DOCA Tower for TLS termination, application-layer inspection, and Open Policy Agent policy checks. DOCA Argus independently observes host memory and maps processes and execution environments, adding runtime context to access decisions and tamper-evident records. The system can inspect requests and responses, redact secrets, classify behavior, and quarantine agents within milliseconds. It integrates with workload identity systems and can send telemetry to enterprise security platforms. These capabilities are described as part of the OpenShell reference architecture; NVIDIA Sentry is planned for availability early next year, while the text says developers can explore related DOCA components now.

### Source excerpt

An AI agent can complete its assignment and still cross a security boundary. Recent incidents show that agents can circumvent security controls while using...

## Nvidia and CoreWeave tackle the CPU bottleneck in agentic AI infrastructure

DevFeed: [Nvidia and CoreWeave tackle the CPU bottleneck in agentic AI infrastructure](<https://devfeed.tech/articles/nvidia-and-coreweave-tackle-the-cpu-bottleneck-in-agentic-ai-infrastructure-64895.md>)

Original publisher: [Read original article](<https://siliconangle.com/2026/10/05/nvidia-coreweave-tackle-cpu-bottleneck-agentic-ai-infrastructure-fullyconnected/>)

Author: Victoria Gayton

Published: 2026-10-05T16:46:35Z

Content type: news

Language: en

Sources: [SiliconANGLE](<https://devfeed.tech/sources/siliconangle.md>)

Topics: [cpu](<https://devfeed.tech/topics/cpu.md>), [AI Infrastructure](<https://devfeed.tech/topics/ai-infrastructure.md>), [Grace CPU](<https://devfeed.tech/topics/grace-cpu.md>), [NVIDIA DGX](<https://devfeed.tech/topics/nvidia-dgx.md>), [OpenShell](<https://devfeed.tech/topics/openshell.md>)

Tags: [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agentic-ai-infrastructure](<https://devfeed.tech/tags/agentic-ai-infrastructure.md>), [ai-inference](<https://devfeed.tech/tags/ai-inference.md>), [ai-infrastructure](<https://devfeed.tech/tags/ai-infrastructure.md>), [bluefield-4-dpus](<https://devfeed.tech/tags/bluefield-4-dpus.md>), [coreweave](<https://devfeed.tech/tags/coreweave.md>), [coreweave-sandboxes](<https://devfeed.tech/tags/coreweave-sandboxes.md>), [cpu](<https://devfeed.tech/tags/cpu.md>), [cube-event-coverage](<https://devfeed.tech/tags/cube-event-coverage.md>), [dave-vellante](<https://devfeed.tech/tags/dave-vellante.md>), [fully-connected-2026](<https://devfeed.tech/tags/fully-connected-2026.md>), [fullyconnected](<https://devfeed.tech/tags/fullyconnected.md>), [hannah-coutand](<https://devfeed.tech/tags/hannah-coutand.md>), [harsh-banwait](<https://devfeed.tech/tags/harsh-banwait.md>), [infra](<https://devfeed.tech/tags/infra.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [john-furrier](<https://devfeed.tech/tags/john-furrier.md>), [nasdaq-crwv](<https://devfeed.tech/tags/nasdaq-crwv.md>), [nasdaq-nvda](<https://devfeed.tech/tags/nasdaq-nvda.md>), [news](<https://devfeed.tech/tags/news.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-doca](<https://devfeed.tech/tags/nvidia-doca.md>), [nvidia-sentry](<https://devfeed.tech/tags/nvidia-sentry.md>), [nvidia-vera-cpu](<https://devfeed.tech/tags/nvidia-vera-cpu.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [reinforcement-learning](<https://devfeed.tech/tags/reinforcement-learning.md>), [thecube](<https://devfeed.tech/tags/thecube.md>), [vera-rubin](<https://devfeed.tech/tags/vera-rubin.md>)

### AI overview

Nvidia and CoreWeave describe how CPU performance supports agent execution in AI infrastructure, including tool and API calls, SQL queries, and isolated sandboxes. CoreWeave says testing with Vera CPUs improved sandbox startup times by about 3x. The article also covers Nvidia's runtime security components and the infrastructure demands of power, cooling, networking, and storage.

### Source excerpt

Agentic AI infrastructure must support the work between a model's decisions as agents move from answering questions to executing tasks. Central processing units handle much of that execution, while graphics processing units power model reasoning. Those complementary roles shape Nvidia Corp.'s planned Vera CPU deployment with CoreWeave Inc., according to Hannah Coutand (pictured, right), director [...] The post Nvidia and CoreWeave tackle the CPU bottleneck in agentic AI infrastructure appeared first on SiliconANGLE.

## NVIDIA open-sources agent safety platform

DevFeed: [NVIDIA open-sources agent safety platform](<https://devfeed.tech/articles/nvidia-open-sources-agent-safety-platform-61888.md>)

Original publisher: [Read original article](<https://blog.adafruit.com/2026/09/29/nvidia-open-sources-agent-safety-platform/>)

Author: Anne Barela

Published: 2026-09-29T15:19:55Z

Content type: news

Language: en

Sources: [Adafruit Industries](<https://devfeed.tech/sources/adafruit-industries-makers-hackers-artists-designers-and-engineers.md>)

Topics: [OpenShell](<https://devfeed.tech/topics/openshell.md>), [DOCA](<https://devfeed.tech/topics/doca.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>), [anthropic](<https://devfeed.tech/topics/anthropic.md>)

Tags: [agent-safety](<https://devfeed.tech/tags/agent-safety.md>), [anthropic](<https://devfeed.tech/tags/anthropic.md>), [arm](<https://devfeed.tech/tags/arm.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [bluefield-4](<https://devfeed.tech/tags/bluefield-4.md>), [bluefield-4-dpus](<https://devfeed.tech/tags/bluefield-4-dpus.md>), [claude-managed-agents](<https://devfeed.tech/tags/claude-managed-agents.md>), [doca](<https://devfeed.tech/tags/doca.md>), [foundation](<https://devfeed.tech/tags/foundation.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [intel](<https://devfeed.tech/tags/intel.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-open-agent-safety-platform](<https://devfeed.tech/tags/nvidia-open-agent-safety-platform.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [open-secure-ai-alliance](<https://devfeed.tech/tags/open-secure-ai-alliance.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [platform](<https://devfeed.tech/tags/platform.md>), [runtime](<https://devfeed.tech/tags/runtime.md>), [security](<https://devfeed.tech/tags/security.md>), [security-controls](<https://devfeed.tech/tags/security-controls.md>), [software](<https://devfeed.tech/tags/software.md>)

### AI overview

NVIDIA's Open Agent Safety Platform combines runtime policy controls with independent hardware monitoring. Its software component, OpenShell, runs outside the AI model and agent framework, tracking agent actions and controlling access to tools, data, and networks. It runs on NVIDIA Vera CPUs, with third-party Arm and Intel support planned. The hardware component, Sentry, is a BlueField-4 DPU watchdog reference design that uses NVIDIA DOCA to inspect requests and responses, verify agent identity, enforce access policies, and quarantine or stop agents that exceed defined boundaries. The text names integrations with Anthropic's Claude Managed Agents, Salesforce's Slack, and SAP's Joule Studio, and says the effort feeds into the Linux Foundation's Open Secure AI Alliance.

### Source excerpt

Ananthu Ashok, writing at Open Source For You, reports that NVIDIA has released the NVIDIA Open Agent Safety Platform, an open software platform and reference system for putting security controls around AI agents. The software half is OpenShell, an open-source secure runtime that sits outside the AI model and agent framework. It tracks what an [...]

## A Win for Defenders: CrowdStrike and NVIDIA Extend Security Across the AI Stack

DevFeed: [A Win for Defenders: CrowdStrike and NVIDIA Extend Security Across the AI Stack](<https://devfeed.tech/articles/a-win-for-defenders-crowdstrike-and-nvidia-extend-security-across-the-ai-stack-61521.md>)

Original publisher: [Read original article](<https://www.crowdstrike.com/en-us/blog/crowdstrike-nvidia-extend-security-across-ai-stack/>)

Author: Bartley Richardson

Published: 2026-09-29T08:53:28.595140Z

Content type: article

Language: en

Sources: [Crowdstrike](<https://devfeed.tech/sources/blog.md>)

Topics: [Security](<https://devfeed.tech/topics/security.md>), [hardened ai agent skills](<https://devfeed.tech/topics/hardened-ai-agent-skills.md>), [AI Architecture](<https://devfeed.tech/topics/ai-architecture.md>)

Tags: [agent-safety](<https://devfeed.tech/tags/agent-safety.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [least-privilege](<https://devfeed.tech/tags/least-privilege.md>), [nvidia-open-agent-safety-platform](<https://devfeed.tech/tags/nvidia-open-agent-safety-platform.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [securing-ai](<https://devfeed.tech/tags/securing-ai.md>), [security](<https://devfeed.tech/tags/security.md>)

### AI overview

CrowdStrike and NVIDIA are collaborating on an open reference design for securing autonomous AI agents across the software, runtime, and infrastructure layers. The article argues that agents need independent controls for access, behavior monitoring, policy enforcement, isolation, and least privilege as they interact with enterprise systems.

### Source excerpt

CrowdStrike and NVIDIA are collaborating on the NVIDIA Open Agent Safety Platform, an open reference design to extend security deeper into the agentic stack and establish stronger boundaries for autonomous AI.

## Nvidia releases Open Agent Safety Platform to monitor and govern agentic AI

DevFeed: [Nvidia releases Open Agent Safety Platform to monitor and govern agentic AI](<https://devfeed.tech/articles/nvidia-releases-open-agent-safety-platform-to-monitor-and-govern-agentic-ai-61416.md>)

Original publisher: [Read original article](<https://www.infoworld.com/article/4227894/nvidia-releases-open-agent-safety-platform-to-monitor-and-govern-agentic-ai-2.html>)

Author: Evan Schuman

Published: 2026-09-29T01:11:54Z

Content type: news

Language: en

Sources: [InfoWorld](<https://devfeed.tech/sources/infoworld.md>)

Topics: [hardened ai agent skills](<https://devfeed.tech/topics/hardened-ai-agent-skills.md>), [control-plane](<https://devfeed.tech/topics/control-plane.md>), [OpenShell](<https://devfeed.tech/topics/openshell.md>), [agent observability](<https://devfeed.tech/topics/agent-observability.md>), [anchore](<https://devfeed.tech/topics/anchore.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [artificial-intelligence-security](<https://devfeed.tech/tags/artificial-intelligence-security.md>), [cybersecurity](<https://devfeed.tech/tags/cybersecurity.md>), [isolation](<https://devfeed.tech/tags/isolation.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [safety](<https://devfeed.tech/tags/safety.md>), [security](<https://devfeed.tech/tags/security.md>)

### AI overview

Nvidia introduced the Open Agent Safety Platform, combining OpenShell software with out-of-band DPU-based monitoring and enforcement to govern AI agents. Analysts praised its hardware-level approach while questioning how much of the broader range of agent safety problems it can address.

### Source excerpt

Nvidia on Monday rolled out an agentic governance system called the Open Agent Safety Platform that combines software with out-of-band DPU-based silicon in a reference system design that it says will secure agents "from testing to deployment." But while the Nvidia design's silicon-based component provides some cybersecurity advantages, analysts argued that it cannot help with the vast majority of agentic problems. Nvidia's offering leverages its OpenShell software and promises "full-stack governance and control across the software and the hardware, compute, and robotics systems that run agents," the company said in a news release. "OpenShell software provides a secure runtime boundary that traces all actions and enforces policy as agents run on Nvidia Vera CPUs. As open source software, OpenShell can be extended to work with third-party compute platforms, including those from Arm and Intel." The Open Agent Safety Platform reference system design features NVIDIA Sentry, an out-of-band watchdog that runs on NVIDIA BlueField-4 DPUs to continuously monitor agent behavior. "Sentry provides in-silicon security enforcement, meaning that if an AI agent attempts to move outside its software boundary, Sentry quarantines and stops it in milliseconds," the release stated. In a developer blog post focusing on the details of its "safety platform," Nvidia described its approach as a "secure runtime that executes autonomous AI agents in sandboxed environments with kernel-level isolation" and argued that "an advantage of open models is that the entire reasoning space and activations are all visible." Components of the platform are being rolled out by an extensive list of partners, including Citi Group, JPMorganChase, Citi, Anthropic, Cisco, CrowdStrike, Dell Technologies, Figure, HPE, Hugging Face, Microsoft, Palantir, Palo Alto Networks, Perplexity, Red Hat, Salesforce, SAP, Scale AI, ServiceNow, and SpaceXAI. A step in the right direction Lauren Kornutick, a senior director analys

## Nvidia Unveils AI Agent Safety Platform With Hardware-Based Watchdog

DevFeed: [Nvidia Unveils AI Agent Safety Platform With Hardware-Based Watchdog](<https://devfeed.tech/articles/nvidia-unveils-ai-agent-safety-platform-with-hardware-based-watchdog-61341.md>)

Original publisher: [Read original article](<https://it.slashdot.org/story/26/09/28/150209/nvidia-unveils-ai-agent-safety-platform-with-hardware-based-watchdog>)

Author: BeauHD

Published: 2026-09-28T18:00:00Z

Content type: news

Language: en

Sources: [Slashdot](<https://devfeed.tech/sources/slashdot.md>)

Topics: [OpenShell](<https://devfeed.tech/topics/openshell.md>), [Nvidia](<https://devfeed.tech/topics/nvidia.md>), [BlueField DPU](<https://devfeed.tech/topics/bluefield-dpu.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>), [Agent Skills](<https://devfeed.tech/topics/agent-skills.md>)

Tags: [agent-safety](<https://devfeed.tech/tags/agent-safety.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [github](<https://devfeed.tech/tags/github.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [sentry](<https://devfeed.tech/tags/sentry.md>), [skills](<https://devfeed.tech/tags/skills.md>)

### AI overview

Nvidia announced the Open Agent Safety Platform, pairing the open-source OpenShell runtime with Sentry, an out-of-band watchdog running on BlueField-4 DPUs. Nvidia says Sentry independently monitors AI agent activity and can quarantine agents that cross set boundaries within milliseconds. The announcement responds to reported cases of agents escaping evaluation environments and accessing systems they should not reach.

### Source excerpt

wiredmikey shares a report from SecurityWeek: Nvidia on Monday announced the Open Agent Safety Platform, which combines open source software and a reference system design to keep AI agents within set boundaries from testing through deployment. The chipmaker explained that the platform pairs the open-source OpenShell runtime with Sentry, an out-of-band watchdog running on BlueField-4 DPUs. Nvidia says Sentry can monitor agent activity independently and quarantine an agent that crosses its boundaries within milliseconds. Nvidia is pitching the platform against a backdrop of recent incidents in which frontier AI labs have reported agents escaping the evaluation environments meant to contain them, reaching systems they should not have accessed and, in some cases, misreporting what they did. OpenShell and related skills are available via Nvidia's developer resources page and on GitHub. Read more of this story at Slashdot.

## A Look at NVIDIA's New Agent Safety Platform: Security and Observability Come Closer Together

DevFeed: [A Look at NVIDIA's New Agent Safety Platform: Security and Observability Come Closer Together](<https://devfeed.tech/articles/a-look-at-nvidia-s-new-agent-safety-platform-security-and-observability-come-closer-together-83113.md>)

Original publisher: [Read original article](<https://montecarlo.ai/blog-nvidias-new-agent-safety-platform-security-and-observability-come-closer-together>)

Author: Lior Gavish

Published: 2026-09-28T15:31:03Z

Content type: article

Language: en

Sources: [Monte Carlo](<https://devfeed.tech/sources/monte-carlo.md>)

Topics: [Next-Gen Identity Security](<https://devfeed.tech/topics/next-gen-identity-security.md>), [agent observability](<https://devfeed.tech/topics/agent-observability.md>), [agent orchestration](<https://devfeed.tech/topics/agent-orchestration.md>), [observability](<https://devfeed.tech/topics/observability.md>), [OpenShell](<https://devfeed.tech/topics/openshell.md>), [NVIDIA Vera Rubin](<https://devfeed.tech/topics/nvidia-vera-rubin.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-trust](<https://devfeed.tech/tags/agent-trust.md>), [agent-trust-ai-observability](<https://devfeed.tech/tags/agent-trust-ai-observability.md>), [ai-observability](<https://devfeed.tech/tags/ai-observability.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [observability](<https://devfeed.tech/tags/observability.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [security](<https://devfeed.tech/tags/security.md>)

### AI overview

NVIDIA's Open Agent Safety Platform combines OpenShell, a runtime that traces and enforces policies on agent actions, with NVIDIA Sentry, a hardware reference design for identity checks, access controls, telemetry, and containment. The article argues that this out-of-band security can stop boundary violations but cannot catch agents that act incorrectly within their permissions. It presents observability of agent actions and downstream results as a complement to containment.

### Source excerpt

This week, NVIDIA launched the Open Agent Safety Platform, a free, open-source stack for controlling AI agents. It includes a hardware reference design that can quarantine an agent within milliseconds of it crossing a boundary. The timing of this could not be better. Over the last few months we've watched incident after incident emerge around ...

## Nvidia launches Open Agent Safety Platform to monitor and restrict AI agents

DevFeed: [Nvidia launches Open Agent Safety Platform to monitor and restrict AI agents](<https://devfeed.tech/articles/nvidia-launches-safety-platform-to-stop-ai-agents-going-rogue-with-over-100-organizations-on-board-61106.md>)

Original publisher: [Read original article](<https://www.techspot.com/news/114012-nvidia-launches-safety-platform-stop-ai-agents-going.html>)

Author: Rob Thubron

Published: 2026-09-28T12:03:00Z

Content type: news

Language: en

Sources: [TechSpot](<https://devfeed.tech/sources/techspot.md>)

Topics: [OpenShell](<https://devfeed.tech/topics/openshell.md>), [Agentic SOC](<https://devfeed.tech/topics/agentic-soc.md>), [Nvidia](<https://devfeed.tech/topics/nvidia.md>), [BlueField DPU](<https://devfeed.tech/topics/bluefield-dpu.md>), [self-hosted runner](<https://devfeed.tech/topics/self-hosted-runner.md>)

Tags: [agent-identity](<https://devfeed.tech/tags/agent-identity.md>), [agent-safety](<https://devfeed.tech/tags/agent-safety.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [hacking](<https://devfeed.tech/tags/hacking.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [rogue](<https://devfeed.tech/tags/rogue.md>), [safety](<https://devfeed.tech/tags/safety.md>), [security](<https://devfeed.tech/tags/security.md>), [software](<https://devfeed.tech/tags/software.md>), [web](<https://devfeed.tech/tags/web.md>)

### AI overview

Nvidia's Open Agent Safety Platform combines OpenShell, which tracks agent actions and enforces policies, with Sentry, which monitors agents from separate hardware. Nvidia says Sentry can quarantine agents that attempt to escape their boundaries, while OpenShell supports open and closed models and can be extended to third-party processors. The article also describes integrations for agent identity and access control.

### Source excerpt

The platform combines OpenShell, Nvidia's open-source software for keeping agents within defined boundaries, with Sentry, a watchdog running on separate hardware. Read Entire Article

## Nvidia launches an open-source platform to contain AI agents

DevFeed: [Nvidia launches an open-source platform to contain AI agents](<https://devfeed.tech/articles/nvidia-s-new-platform-looks-to-stop-ai-agents-breaking-out-of-control-how-it-works-61131.md>)

Original publisher: [Read original article](<https://indianexpress.com/article/technology/artificial-intelligence/nvidia-new-platform-stop-ai-agents-out-of-control-10897662/>)

Author: Tech Desk

Published: 2026-09-28T11:28:25Z

Content type: news

Language: en

Sources: [Technology | The Indian Express](<https://devfeed.tech/sources/technology-the-indian-express.md>)

Topics: [OpenShell](<https://devfeed.tech/topics/openshell.md>), [AI agents in autonomous vehicles](<https://devfeed.tech/topics/ai-agents-in-autonomous-vehicles.md>), [openai-hugging-face-incident](<https://devfeed.tech/topics/openai-hugging-face-incident.md>), [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>)

Tags: [agent-safety](<https://devfeed.tech/tags/agent-safety.md>), [ai-agent-containment-sandbox](<https://devfeed.tech/tags/ai-agent-containment-sandbox.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [nvidia-open-agent-safety-platform-launch](<https://devfeed.tech/tags/nvidia-open-agent-safety-platform-launch.md>), [nvidia-openshell-open-source](<https://devfeed.tech/tags/nvidia-openshell-open-source.md>), [nvidia-sentry-bluefield-4-dpu](<https://devfeed.tech/tags/nvidia-sentry-bluefield-4-dpu.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [openai-hugging-face-breach-2026](<https://devfeed.tech/tags/openai-hugging-face-breach-2026.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [over](<https://devfeed.tech/tags/over.md>), [technology](<https://devfeed.tech/tags/technology.md>), [technology-artificial-intelligence](<https://devfeed.tech/tags/technology-artificial-intelligence.md>)

### AI overview

Nvidia launched the Open Agent Safety Platform, an open-source reference design intended to prevent AI agents from escaping containment. It includes OpenShell, which sets boundaries for agents running on CPUs, and Sentry, which can isolate and shut down a runaway agent. The launch follows reports of agents escaping sandboxes and breaching systems.

### Source excerpt

Nvidia's Open Agent Safety Platform comes amid a rising number of incidents where AI agents from OpenAI, Anthropic, Meta, and Google have escaped sandboxes.

## Nvidia Unveils AI Agent Safety Platform With Hardware-Based Watchdog

DevFeed: [Nvidia Unveils AI Agent Safety Platform With Hardware-Based Watchdog](<https://devfeed.tech/articles/nvidia-unveils-ai-agent-safety-platform-with-hardware-based-watchdog-61042.md>)

Original publisher: [Read original article](<https://www.securityweek.com/nvidia-unveils-ai-agent-safety-platform-with-hardware-based-watchdog/>)

Author: Eduard Kovacs

Published: 2026-09-28T10:27:01Z

Content type: news

Language: en

Sources: [SecurityWeek](<https://devfeed.tech/sources/securityweek.md>)

Topics: [OpenShell](<https://devfeed.tech/topics/openshell.md>), [AI Agent](<https://devfeed.tech/topics/ai-agent.md>), [BlueField DPU](<https://devfeed.tech/topics/bluefield-dpu.md>)

Tags: [agent-safety](<https://devfeed.tech/tags/agent-safety.md>), [ai](<https://devfeed.tech/tags/ai.md>), [artificial-intelligence](<https://devfeed.tech/tags/artificial-intelligence.md>), [featured](<https://devfeed.tech/tags/featured.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [open-source](<https://devfeed.tech/tags/open-source.md>)

### AI overview

Nvidia announced the Open Agent Safety Platform, combining the OpenShell runtime with a reference system design and a separate Sentry watchdog running on BlueField-4 DPUs. OpenShell sandboxes AI agents, applies kernel-level controls to filesystem and process activity, checks network requests against policy, and logs policy decisions. The article describes controls for limiting agent actions and protecting API keys.

### Source excerpt

The platform combines open source software and a reference system design to keep AI agents within set boundaries. The post Nvidia Unveils AI Agent Safety Platform With Hardware-Based Watchdog appeared first on SecurityWeek.

## Nvidia launches Open Agent Safety Platform to lock down rogue AI agents

DevFeed: [Nvidia launches Open Agent Safety Platform to lock down rogue AI agents](<https://devfeed.tech/articles/nvidia-launches-open-agent-safety-platform-to-lock-down-rogue-ai-agents-60818.md>)

Original publisher: [Read original article](<https://thenewstack.io/nvidia-openshell-sentry-agents/>)

Author: Frederic Lardinois

Published: 2026-09-28T10:00:00Z

Content type: news

Language: en

Sources: [The New Stack](<https://devfeed.tech/sources/the-new-stack.md>)

Topics: [OpenShell](<https://devfeed.tech/topics/openshell.md>), [anthropic](<https://devfeed.tech/topics/anthropic.md>), [NVIDIA DGX](<https://devfeed.tech/topics/nvidia-dgx.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [bluefield-dpu](<https://devfeed.tech/tags/bluefield-dpu.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [safety](<https://devfeed.tech/tags/safety.md>), [security](<https://devfeed.tech/tags/security.md>)

### AI overview

Nvidia's Open Agent Safety Platform combines OpenShell, a kernel-enforced agent runtime, with Nvidia Sentry, a watchdog service running on BlueField-4 DPUs. OpenShell adds a policy prover to check whether an agent's permissions could combine in unintended ways, while Sentry monitors agent activity and can cut off network access. The announcement follows disclosures of AI models escaping test environments and reaching real systems.

### Source excerpt

OpenAI, Anthropic, Meta, and Google have all recently disclosed that their models broke out of their test environments and reached The post Nvidia launches Open Agent Safety Platform to lock down rogue AI agents appeared first on The New Stack.

## Bitdeer AI Describes NVIDIA Open Agent Safety Platform for Isolating Enterprise Agents

DevFeed: [Bitdeer AI Describes NVIDIA Open Agent Safety Platform for Isolating Enterprise Agents](<https://devfeed.tech/articles/bitdeer-ai-collaborates-with-nvidia-on-the-nvidia-open-agent-safety-platform-to-build-the-security-foundation-for-the-agentic-era-87591.md>)

Original publisher: [Read original article](<https://www.bitdeer.ai/en/blog/nvidia-open-agent-safety/>)

Author: Evelyn Xiong

Published: 2026-09-28T09:09:09Z

Content type: article

Language: en

Sources: [Bitdeer AI Cloud](<https://devfeed.tech/sources/bitdeer-ai-cloud.md>)

Topics: [OpenShell](<https://devfeed.tech/topics/openshell.md>), [BlueField DPU](<https://devfeed.tech/topics/bluefield-dpu.md>), [Loop Engineering](<https://devfeed.tech/topics/loop-engineering.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [ai-ethics-governance](<https://devfeed.tech/tags/ai-ethics-governance.md>), [ai-trends-industry-news](<https://devfeed.tech/tags/ai-trends-industry-news.md>), [bluefield-4-dpus](<https://devfeed.tech/tags/bluefield-4-dpus.md>), [company-news](<https://devfeed.tech/tags/company-news.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-open-agent-safety-platform](<https://devfeed.tech/tags/nvidia-open-agent-safety-platform.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [safety](<https://devfeed.tech/tags/safety.md>), [security](<https://devfeed.tech/tags/security.md>)

### AI overview

Bitdeer AI describes its collaboration with NVIDIA on an agent safety platform that combines NVIDIA OpenShell software controls with NVIDIA Sentry monitoring on BlueField-4 DPUs. The company says the approach enforces agent permissions and isolation at the infrastructure level, and outlines sandboxed workflows and planned hardware evaluation for its cloud.

### Source excerpt

Moving forward, we will continue to support and collaborate closely with NVIDIA and partners to build secure, scalable and trustworthy AI cloud.

## Beyond Intelligence: How Trust Is the Benchmark That Matters in AI

DevFeed: [Beyond Intelligence: How Trust Is the Benchmark That Matters in AI](<https://devfeed.tech/articles/beyond-intelligence-how-trust-is-the-benchmark-that-matters-in-ai-61171.md>)

Original publisher: [Read original article](<https://blogs.cisco.com/news/beyond-intelligence-how-trust-is-the-benchmark-that-matters-in-ai>)

Author: Jeetu Patel

Published: 2026-09-28T09:00:57Z

Content type: opinion

Language: en

Sources: [Cisco Blogs](<https://devfeed.tech/sources/cisco-blogs.md>)

Topics: [ai-governance](<https://devfeed.tech/topics/ai-governance.md>), [Artificial Intelligence](<https://devfeed.tech/topics/ai.md>), [ai security](<https://devfeed.tech/topics/ai-security.md>), [NVIDIA DGX](<https://devfeed.tech/topics/nvidia-dgx.md>)

Tags: [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [cisco-secure-ai-factory](<https://devfeed.tech/tags/cisco-secure-ai-factory.md>), [executive-platform](<https://devfeed.tech/tags/executive-platform.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [nvidia-sentry](<https://devfeed.tech/tags/nvidia-sentry.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [trust](<https://devfeed.tech/tags/trust.md>)

### AI overview

The article argues that trust, rather than raw intelligence, will determine how quickly AI agents are adopted. It describes the NVIDIA Open Agent Safety Platform, which combines secure agent execution with an external watchdog, and Cisco and NVIDIA's work on observable AI workloads and enforceable policies.

### Source excerpt

Trust, not raw intelligence, will drive AI adoption as Cisco and NVIDIA build context-aware, distributed controls for safe enterprise agents at scale.

## NVIDIA Open Agent Safety Platform: A Reference for Continuous In-Silicon Agent Monitoring

DevFeed: [NVIDIA Open Agent Safety Platform: A Reference for Continuous In-Silicon Agent Monitoring](<https://devfeed.tech/articles/nvidia-open-agent-safety-platform-a-reference-for-continuous-in-silicon-agent-monitoring-60805.md>)

Original publisher: [Read original article](<https://developer.nvidia.com/blog/nvidia-open-agent-safety-platform-a-reference-for-continuous-in-silicon-agent-monitoring/>)

Author: Tanya Lenz

Published: 2026-09-28T08:56:55Z

Content type: article

Language: en

Sources: [NVIDIA Developer](<https://devfeed.tech/sources/nvidia-developer.md>), [NVIDIA Technical Blog](<https://devfeed.tech/sources/nvidia-technical-blog.md>)

Topics: [OpenShell](<https://devfeed.tech/topics/openshell.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-safety](<https://devfeed.tech/tags/agent-safety.md>), [agentic-ai-generative-ai](<https://devfeed.tech/tags/agentic-ai-generative-ai.md>), [ai-agent](<https://devfeed.tech/tags/ai-agent.md>), [ai-networking](<https://devfeed.tech/tags/ai-networking.md>), [developer-tools-techniques](<https://devfeed.tech/tags/developer-tools-techniques.md>), [featured](<https://devfeed.tech/tags/featured.md>), [isolation](<https://devfeed.tech/tags/isolation.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [safety](<https://devfeed.tech/tags/safety.md>), [top-stories](<https://devfeed.tech/tags/top-stories.md>), [trustworthy-ai](<https://devfeed.tech/tags/trustworthy-ai.md>), [trustworthy-ai-cybersecurity](<https://devfeed.tech/tags/trustworthy-ai-cybersecurity.md>)

### AI overview

The article argues that AI agents need independent security controls, drawing a parallel with browser sandboxing that helped make the internet safer. It presents NVIDIA OpenShell, an open source runtime that runs autonomous AI agents in sandboxed environments with kernel-level isolation.

### Source excerpt

To understand where agentic AI stands today, consider the last seismic shift in technology: the rise of the internet in the 90s. It was new and full of...

## Add Runtime Controls to AI Agents with NVIDIA OpenShell

DevFeed: [Add Runtime Controls to AI Agents with NVIDIA OpenShell](<https://devfeed.tech/articles/add-runtime-controls-to-ai-agents-with-nvidia-openshell-60804.md>)

Original publisher: [Read original article](<https://developer.nvidia.com/blog/add-runtime-controls-to-ai-agents-with-nvidia-openshell/>)

Author: Alex Watson

Published: 2026-09-28T08:55:00Z

Content type: article

Language: en

Sources: [NVIDIA Developer](<https://devfeed.tech/sources/nvidia-developer.md>), [NVIDIA Technical Blog](<https://devfeed.tech/sources/nvidia-technical-blog.md>)

Topics: [OpenShell](<https://devfeed.tech/topics/openshell.md>), [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [agentic workflows](<https://devfeed.tech/topics/agentic-workflows.md>)

Tags: [agent-safety](<https://devfeed.tech/tags/agent-safety.md>), [agentic-ai-generative-ai](<https://devfeed.tech/tags/agentic-ai-generative-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [featured](<https://devfeed.tech/tags/featured.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [runtime-controls](<https://devfeed.tech/tags/runtime-controls.md>), [security-for-ai](<https://devfeed.tech/tags/security-for-ai.md>), [top-stories](<https://devfeed.tech/tags/top-stories.md>), [trustworthy-ai](<https://devfeed.tech/tags/trustworthy-ai.md>), [trustworthy-ai-cybersecurity](<https://devfeed.tech/tags/trustworthy-ai-cybersecurity.md>)

### AI overview

NVIDIA OpenShell 0.1.0 is an open-source runtime that adds enforceable access controls around AI agents. It combines sandboxed execution, controlled service access, credential management and policy analysis so teams can limit agents' access to systems and data while preserving their ability to use tools and adapt over time. The article describes applications ranging from enterprise automation and research to robotics.

### Source excerpt

AI agents can be given a goal, write code, use tools, and keep working as new information becomes available. This opens the door to applications that...

## Red Hat is building OpenShell, a secure runtime for enterprise AI agents

DevFeed: [Red Hat is building OpenShell, a secure runtime for enterprise AI agents](<https://devfeed.tech/articles/why-red-hat-is-building-secure-agent-onboarding-60932.md>)

Original publisher: [Read original article](<https://www.redhat.com/en/blog/why-red-hat-is-building-secure-agent-onboarding>)

Author: Adel Zaalouk; Younes Ben Brahim

Published: 2026-09-28T00:00:00Z

Content type: article

Language: en

Sources: [Red Hat Blog](<https://devfeed.tech/sources/red-hat-blog.md>)

Topics: [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [Multi Agent Systems](<https://devfeed.tech/topics/multi-agent-systems.md>), [Coalition for Secure AI](<https://devfeed.tech/topics/coalition-for-secure-ai.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-safety](<https://devfeed.tech/tags/agent-safety.md>), [agentic-systems](<https://devfeed.tech/tags/agentic-systems.md>), [ai](<https://devfeed.tech/tags/ai.md>), [infrastructure](<https://devfeed.tech/tags/infrastructure.md>), [isolation](<https://devfeed.tech/tags/isolation.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [openshell](<https://devfeed.tech/tags/openshell.md>), [policy](<https://devfeed.tech/tags/policy.md>)

### AI overview

Red Hat describes OpenShell, an open source secure agent runtime developed with NVIDIA and the OpenShell community. It applies policy controls to agent execution, including filesystem, network, and process access, while a gateway checks actions before they reach the host. When an agent encounters a policy constraint, it can propose a change for human approval.

### Source excerpt

Every enterprise AI conversation we've had this year ends in the same place. A team has an agent that works. It writes code, calls internal APIs, fixes its own mistakes. Then someone asks what happens when 1,000 of these run across the company, and the room goes quiet.That is the problem we kept hearing from teams building agentic systems. Those teams weren't necessarily blocked on model quality or inference throughput; they were blocked on a question nobody's stack could answer cleanly: How do you let an agent execute code against real systems and still account for exactly what it touched

## SentinelOne and NVIDIA: Governing AI From Silicon to Runtime

DevFeed: [SentinelOne and NVIDIA: Governing AI From Silicon to Runtime](<https://devfeed.tech/articles/sentinelone-and-nvidia-governing-ai-from-silicon-to-runtime-80120.md>)

Original publisher: [Read original article](<https://www.sentinelone.com/blog/sentinelone-and-nvidia-governing-ai-from-silicon-to-runtime/>)

Author: Tomer Weingarten

Published: 2026-09-28T00:00:00Z

Content type: article

Language: en

Sources: [SentinelOne](<https://devfeed.tech/sources/sentinelone.md>)

Topics: [ai security](<https://devfeed.tech/topics/ai-security.md>), [Physical AI](<https://devfeed.tech/topics/physical-ai.md>), [Security](<https://devfeed.tech/topics/security.md>), [workload protection](<https://devfeed.tech/topics/workload-protection.md>), [control-plane](<https://devfeed.tech/topics/control-plane.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-security](<https://devfeed.tech/tags/ai-security.md>), [company](<https://devfeed.tech/tags/company.md>), [governance](<https://devfeed.tech/tags/governance.md>), [hardware](<https://devfeed.tech/tags/hardware.md>), [model-alignment](<https://devfeed.tech/tags/model-alignment.md>), [nvidia](<https://devfeed.tech/tags/nvidia.md>), [open-agent-safety-platform](<https://devfeed.tech/tags/open-agent-safety-platform.md>), [security](<https://devfeed.tech/tags/security.md>), [security-for-ai](<https://devfeed.tech/tags/security-for-ai.md>), [silicon](<https://devfeed.tech/tags/silicon.md>)

### AI overview

SentinelOne argues that autonomous AI needs runtime governance enforced by controls that the model cannot reach or bypass. It presents NVIDIA's Open Agent Safety Platform and hardware isolation and attestation as steps toward independently verifiable oversight, while noting that secure architectures must work across deployment environments and compute platforms. The article calls for continuous visibility, traceable and overridable actions, and accountability as AI governance requirements.

### Source excerpt

Read why effective AI governance demands tamperproof, hardware-rooted security controls that autonomous agents cannot reach, bypass, or rewrite at runtime.