Agent Security
2 stories in this blend
The Neuron·Oct 22Nvidia announces OpenShell security framework for autonomous agents
Nvidia unveiled an architecture specification called OpenShell aimed at restricting and auditing what AI agents can access on host systems. Early industry adopters of the reference design include Microsoft, Anthropic, and OpenAI.
There's An AI For That·Sep 26AxioRank
AxioRank inspects API calls made by software agents to block leaked credentials, destructive shell commands, and prompt injection attempts. It fits security teams that need control and visibility over autonomous AI agent deployments.
