CogNEXUS insights

Blog

Field notes on AI governance, the EU AI Act, agent safety, and ArtzAIn, the model-agnostic agent control plane.

OpenShell keeps the agent inside the fence. ArtzAIn decides, and proves.

NVIDIA’s Open Agent Safety Platform advances agent containment. ArtzAIn complements it as the model-agnostic agent control plane: Policy before the action, Proof after.

Read the post

One question, one verdict, one receipt: a tour of ArtzAIn

What ArtzAIn does today, what it plugs into, and four ways to run it, from a free artzain SDK on your laptop to an air-gapped deployment. Free guards, signed policy, offline-verifiable evidence packs, discovery across twelve sources, and 23 live connectors.

Read the post

Achieving EU AI Act compliance with ArtzAIn

How the control plane turns "we think we're compliant" into a signed, offline-verifiable evidence pack, a full year before the deadline. A walk through Articles 12, 9, and 10, with runnable examples from the artzain SDK.

Read the post