According to a recent LinkedIn post from Nirmata, the company is promoting a CNCF blog guide focused on securing artificial intelligence agents running on Kubernetes. The post highlights risks tied to agents’ ability to execute shell commands, install packages, access files and use credentials.
The guide reportedly covers agent sandboxing with OpenShell, request-level network restrictions and the use of credential placeholders to prevent direct secret exposure. It also includes seven Kyverno policies addressing pods, container images, network access and credentials, along with a demonstration repository for testing in a Kubernetes cluster.
For Nirmata, the material may support positioning around cloud-native security and policy management as enterprises deploy more autonomous software agents. Greater attention to workload isolation and access controls could expand demand for governance tools in Kubernetes environments, although the post does not provide customer, revenue or adoption data.
The focus on practical policies and an available demo may help encourage evaluation by platform engineering and security teams. However, the potential financial impact remains difficult to assess without evidence of commercial conversion, partnerships or changes in the company’s operating outlook.


