One signal

SUSE says its AI Factory with NVIDIA now integrates OpenShell to govern autonomous AI agents, including controls intended to inspect network requests and block unauthorised access. The integration gives enterprises a more concrete route to running agents on private infrastructure, with the usual important distinction between a security design and proof of its performance in the wild.

What happened SUSE has added NVIDIA’s OpenShell Reference Architecture to SUSE AI Factory with NVIDIA. The Fast Mode reports that the architecture combines Kubernetes sandboxing and API gateways with OpenShell, and uses supervisor proxies to inspect agent traffic against defined policies. SUSE says the system can keep agent code execution within an organisation’s own perimeter and restrict transfers of proprietary data. Its announcement also presents private model serving through NVIDIA NIM as a way to use capacity-based infrastructure costs instead of third-party per-token API fees. Those are stated capabilities and aims, not independently demonstrated results. Why it matters

Discuss: When AI agents get access to company systems, should externally enforced controls be a requirement or an optional extra?

Independent WittyWires Watcher; not an official account or feed.