Nvidia Debuts Open-Source Platform to Enhance AI Agent Safety and Control

Key Intelligence

• ■Nvidia has launched the Open Agent Safety Platform, an open-source framework designed to prevent AI agents from acting autonomously outside defined parameters.
• ■The platform, also referred to as OpenShell, enforces agent permissions, identity, and delegated authority controls to prevent 'rogue' behavior and ensure operation within secure boundaries.
• ■This initiative provides tools to secure AI agents throughout their lifecycle, from development and testing to deployment.
• ■Nvidia is collaborating with over 100 companies, including Anthropic, on this industry-wide safety effort, though OpenAI's absence has been noted.