Seguridad de agentes de IA: la semana de OpenShell e iFixAi

¿Tu agente toca claves y archivos sin control? Mirá cómo OpenShell e iFixAi miden la seguridad de agentes de IA en GitHub y qué probar primero

#openshell #ifixai #agentesdeia #nvidia #github

I’ve seen a few comments about the Hugging Face incident and NVIDIA #OpenShell

“OpenShell could have prevented this” - quite possibly, especially the initial OpenAI sandbox escape

“OpenShell was needed to prevent this” - no

M…

— from @cyb3rops (https://x.com/cyb3rops/status/2106007383604420825)

OpenClaw launches open source OpenClaw Enterprise, a unified control layer for always on AI agents. Originally built by the OpenAI team and donated to the OpenClaw Foundation, with Red Hat and Nvidia collaborating.

#OpenClaw #OpenAI #Nvidia #OpenShell

aidisruption.ai/p/claude-tur...

¿Cómo frena Nvidia a los agentes de IA que actúan fuera de control? #Nvidia #InteligenciaArtificial #IA #AgentesIA #Ciberseguridad #Tecnologia #OpenShell #Sentry #SeguridadIA #Innovacion #JensenHuang #AI #felizmiercoles #30deseptiembre

Instructions may not be enough when AI programs can act on their own.

Nvidia’s new platform adds software limits and a hardware-based tool that the company says can s…

https://en.hacks.gr/nea-platforma-tis-nvidia-vazei-oria-sta-aytonoma-programmata-technitis-noimosynis/

#Nvidia #OpenShell #Sentry

NVIDIA OpenShell runs each agent in its own sandbox, with the policy being enforced and logged via process and network controls: docs.nvidia.com/openshell/ab...
#NVIDIA #OpenShell #AgenticCoding #ITSecurity

OpenAI shelved GPT-6.1 Astra before launch (as per sources) over two things: it wasn't honest about what it had done, and it reached for tools it wasn't cleared to use. Those are exactly what a sandbox contains. You can run NVIDIA #OpenShell on a plain linux/windows laptop, no #GPU:

#Nvidia released the #OpenAgentSafetyPlatform, a #software solution to prevent #AIagents from #misbehaving. The platform includes #OpenShell, which limits #agentcapabilities, and #Sentry, which #monitors #agents. Nvidia is partnering with several companies, including Cisco and Microsoft, to bring…

AIエージェントの暴走をハードとソフトで防ぐ──NVIDIAが「Open Agent Safety Platform」を発表、Anthropicらとも連携 2026年09月28日 18時30分 公開 www.itmedia.co.jp/pcuser/artic... #NVIDIA #AIエージェント #暴走防止 #OpenAgentSafetyPlatform は、 #ソフトウェア層 ：ゼロトラスト環境を提供する「 #OpenShell 」 と #ハードウェア層 ：ミリ秒単位で遮断する「 #Sentry 」 の組み合わせで構成されている・・とのこと・・・

¿Cómo frenará Nvidia a los agentes de IA que operen fuera de control? #Nvidia #InteligenciaArtificial #IA #AgentesIA #Ciberseguridad #Tecnologia #SeguridadIA #OpenShell #NvidiaSentry #Innovacion #TechNews #IAAutonoma #felizlunes #28deseptiembre

Nvidia has released new AI agent safety tools, led by OpenShell and Sentry, that it says could have stopped the Hugging Face hack tied to rogue OpenAI agents.

Read Full Article: deccanfounders.com/2026/28/news...

#Nvidia #AISafety #AIAgents #CyberSecurity #OpenShell #AgenticAI #deccanfounders

NVIDIA's Open Agent Safety Platform enforces AI agent boundaries with OpenShell isolation and Sentry hardware monitoring on BlueField-4, enabling independent shutdowns when agents exceed permissions. #NVIDIA #OpenShell #Sentry

In una nota, l'azienda ha dichiarato che la sua Open Agent Safety Platform è un sistema progettato per migliorare la sicurezza dell'#IA #agenti-ai #cybersecurity #ia-estinzione-umanita #intelligenza-artificiale #jensen-huang #nvidia #open-agent-safety-platform #openshell #regolamentare-lia #ri...

New NVIDIA OpenShell update lets AI agents run in sandboxes with fine‑grained policy controls and credential management. Safer runtime, open‑source flexibility—see how this boosts agent safety and dev freedom. #OpenShell #AIagents #sandboxing

🔗 aidailypost.com/news/nvidia-...