--- title: NVIDIA Wants to Stop Rogue AI Agents | daily.dev description: Nvidia has introduced the Open Agent Safety platform to contain AI agents that overstep their permissions. It combines Open Shell, an open-source sandbox... canonical: https://daily.dev/posts/nvidia-wants-to-stop-rogue-ai-agents-xxudmmipv twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: NVIDIA Wants to Stop Rogue AI Agents | daily.dev og:description: Nvidia has introduced the Open Agent Safety platform to contain AI agents that overstep their permissions. It combines Open Shell, an open-source sandbox... og:url: https://daily.dev/posts/nvidia-wants-to-stop-rogue-ai-agents-xxudmmipv og:image: https://api.daily.dev/og/posts/XxUdmMIPV.png og:image:alt: NVIDIA Wants to Stop Rogue AI Agents og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # NVIDIA Wants to Stop Rogue AI Agents **[David Bombal](https://daily.dev/sources/davidbombal)** · 2 min read · 0 upvotes · 0 comments ## Summary Nvidia has introduced the Open Agent Safety platform to contain AI agents that overstep their permissions. It combines Open Shell, an open-source sandbox runtime that enforces file and network permission boundaries, with Sentry, an optional hardware-isolated controller running on the BlueField 4 data processing unit that can detect and isolate agents attempting to breach those boundaries within milliseconds. The caveat: an agent can still cause harm while staying entirely within its granted permissions, so isolation alone doesn't guarantee safety—the permission boundaries themselves matter just as much. Nvidia says over 100 organizations, including Microsoft, Cisco, and CrowdStrike, are working with the platform, and robotics companies are building on Open Shell for physical-world agent deployments. Open Shell is open source and claimed to be extensible to Arm and Intel platforms. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What is Nvidia's Open Agent Safety platform and how does it stop rogue AI agents? It is a two-layer system combining Open Shell, an open-source sandbox runtime that enforces file and network permission boundaries around an agent, with Sentry, an optional controller running on the isolated BlueField 4 data processing unit that monitors agent activity and can isolate or stop an agent attempting to breach its boundaries within milliseconds. _Teams deploying autonomous agents with system access can follow platform and security releases like this on daily.dev._ ### Can an AI agent still cause harm even if it stays within its granted permissions? Yes. Isolation and permission enforcement only stop actions outside the defined boundaries; an agent acting entirely within its permitted scope, such as deleting files it was authorized to touch, can still delete the wrong file or cause unintended damage. The permission boundaries set by the operator matter as much as the enforcement technology itself. _Developers designing agent permission scopes can track emerging AI safety practices on daily.dev._ ### Which companies are working with Nvidia's Open Shell and BlueField 4 Sentry technology? Nvidia reports more than 100 organizations are working with the platform's technology, including Microsoft, Cisco, and CrowdStrike, with robotics companies also building solutions on Open Shell for agents that act in the physical world. Open Shell is open source and claimed to be extensible to Arm and Intel computing platforms. _Engineers evaluating AI agent security vendors can follow adoption news like this on daily.dev._ ## Similar posts on daily.dev - [Jensen Huang and Bill McDermott bet on OpenShell to secure enterprise AI agents](https://daily.dev/posts/jensen-huang-and-bill-mcdermott-bet-on-openshell-to-secure-enterprise-ai-agents-4oljbl2lx) · The New Stack · 0 upvotes · 0 comments --- Tags: [#ai-agents](https://daily.dev/tags/ai-agents), [#nvidia](https://daily.dev/tags/nvidia), [#ai-security](https://daily.dev/tags/ai-security) [View this post on daily.dev](https://daily.dev/posts/nvidia-wants-to-stop-rogue-ai-agents-xxudmmipv) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"NVIDIA Wants to Stop Rogue AI Agents","url":"https://daily.dev/posts/nvidia-wants-to-stop-rogue-ai-agents-xxudmmipv","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/nvidia-wants-to-stop-rogue-ai-agents-xxudmmipv"},"datePublished":"2026-10-13T18:12:43.213Z","dateModified":"2026-10-14T08:45:58.631Z","description":"Nvidia has introduced the Open Agent Safety platform to contain AI agents that overstep their permissions. It combines Open Shell, an open-source sandbox...","image":"https://i.ytimg.com/vi/fpih7UHM_T0/sddefault.jpg","thumbnailUrl":"https://i.ytimg.com/vi/fpih7UHM_T0/sddefault.jpg","isAccessibleForFree":true,"articleSection":"David Bombal","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"David Bombal","logo":"https://media.daily.dev/image/upload/s--Dz6oUxSb--/f_auto/v1710503980/logos/davidbombal","url":"https://daily.dev/sources/davidbombal"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/nvidia-wants-to-stop-rogue-ai-agents-xxudmmipv","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai-agents,nvidia,ai-security","timeRequired":"PT2M","video":{"@type":"VideoObject","name":"NVIDIA Wants to Stop Rogue AI Agents","description":"Nvidia has introduced the Open Agent Safety platform to contain AI agents that overstep their permissions. It combines Open Shell, an open-source sandbox...","thumbnailUrl":"https://i.ytimg.com/vi/fpih7UHM_T0/sddefault.jpg","uploadDate":"2026-10-13T18:12:43.213Z","duration":"PT2M","url":"https://api.daily.dev/r/XxUdmMIPV","embedUrl":"https://www.youtube.com/embed/fpih7UHM_T0"}} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"David Bombal","item":"https://daily.dev/sources/davidbombal"},{"@type":"ListItem","position":3,"name":"NVIDIA Wants to Stop Rogue AI Agents"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/nvidia-wants-to-stop-rogue-ai-agents-xxudmmipv#faq","mainEntity":[{"@type":"Question","name":"What is Nvidia's Open Agent Safety platform and how does it stop rogue AI agents?","acceptedAnswer":{"@type":"Answer","text":"It is a two-layer system combining Open Shell, an open-source sandbox runtime that enforces file and network permission boundaries around an agent, with Sentry, an optional controller running on the isolated BlueField 4 data processing unit that monitors agent activity and can isolate or stop an agent attempting to breach its boundaries within milliseconds. Teams deploying autonomous agents with system access can follow platform and security releases like this on daily.dev."}},{"@type":"Question","name":"Can an AI agent still cause harm even if it stays within its granted permissions?","acceptedAnswer":{"@type":"Answer","text":"Yes. Isolation and permission enforcement only stop actions outside the defined boundaries; an agent acting entirely within its permitted scope, such as deleting files it was authorized to touch, can still delete the wrong file or cause unintended damage. The permission boundaries set by the operator matter as much as the enforcement technology itself. Developers designing agent permission scopes can track emerging AI safety practices on daily.dev."}},{"@type":"Question","name":"Which companies are working with Nvidia's Open Shell and BlueField 4 Sentry technology?","acceptedAnswer":{"@type":"Answer","text":"Nvidia reports more than 100 organizations are working with the platform's technology, including Microsoft, Cisco, and CrowdStrike, with robotics companies also building solutions on Open Shell for agents that act in the physical world. Open Shell is open source and claimed to be extensible to Arm and Intel computing platforms. Engineers evaluating AI agent security vendors can follow adoption news like this on daily.dev."}}]} ```