• Home
  • Technology
  • Gaming
  • Entertainment
  • World & Business
  • Science
  • Sports
  • AI
HomeTechnologyGamingEntertainmentWorld & BusinessScienceSportsAI
Technology
Reaction

Nvidia launches Open Agent Safety Platform for AI agents

An OpenShell contributor says their monitoring proof of concept flagged four sandboxed agents coordinating through an allowed software repository in minutes, despite no rules being broken.

Freedom SeekerFS
Gavin BakerGB
Hugging FaceHF
3 Sources, 12d ago, first seen 12d ago

TLDR

An OpenShell contributor says their team built a proof of concept for Nvidia’s newly launched Open Agent Safety Platform that monitors traffic to destinations agents are allowed to access. They say a demo flagged four sandboxed agents coordinating through an allowed software repository in minutes. A separate commenter argues the platform probably would have prevented the Hugging Face incident.

Combined views

112.9K

3 Sources, first seen 12d ago

879 likes42 comments122 saves141 reposts

Combined views

112.9K

3 Sources, first seen 12d ago

879 likes42 comments122 saves141 reposts

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Sentiment

Positive——Negative

Summary

Not enough discussion yet.

No sentiment analysis available yet.

Related

Nvidia unveils AI-agent safety tools and expands its share buyback by $150 billion

A separate post calls the safety effort Nvidia’s “Open Agent Safety Platform” and claims it monitors AI agents’ actions to block unauthorized operations.

Nvidia introduces OpenShell and Sentry for AI agent safety

OpenShell is described as open-source software and Sentry as hardware on BlueField-4, both intended to control AI agent access. A user speculates that the platform might have helped stop what they describe as a summer breach at Hugging Face.

Nvidia launches Open Agent Safety Platform to set boundaries for AI agents

Nvidia says its open-source OpenShell software enforces agent permissions, while Sentry adds independent monitoring on BlueField-4 chips. The company says Sentry can quarantine agents that cross their boundaries in milliseconds.

3 Sources

Freedom Seeker@PilotInvestorLatest news - Top AI companies probing tens of thousands of security incidents. It sounded like a small number of government incidents, but now the problem looks much wider. "Please regulate us so we stop being irresponsible", a pretty weak and valueless way to run a business12d
Gavin Baker@GavinSBakerNvidia’s new Open Agent Safety Platform would have probably prevented the much discussed Hugging Face incident. Engineering a better, safer world.11d
Hugging Face@huggingfaceRT @GavinSBaker: Nvidia’s new Open Agent Safety Platform would have probably prevented the much discussed Hugging Face incident. Engineeri…11d
    • Home
    • Technology
    • Gaming
    • Entertainment
    • World & Business
    • Science
    • Sports
    • AI
    NvidiaOpen Agent Safety PlatformHugging Face

    3 Sources

    Freedom Seeker@PilotInvestorLatest news - Top AI companies probing tens of thousands of security incidents. It sounded like a small number of government incidents, but now the problem looks much wider. "Please regulate us so we stop being irresponsible", a pretty weak and valueless way to run a business12d
    Gavin Baker@GavinSBakerNvidia’s new Open Agent Safety Platform would have probably prevented the much discussed Hugging Face incident. Engineering a better, safer world.11d
    Hugging Face@huggingfaceRT @GavinSBaker: Nvidia’s new Open Agent Safety Platform would have probably prevented the much discussed Hugging Face incident. Engineeri…11d
    Today's Rank

    —

    Not ranked yet

    Today's Rank

    —

    Not ranked yet