The New Multi-Tenant Challenge: Securing AI Agents in Cloud-Native Infrastructure
AI agents run untrusted code. Here’s how to secure them using isolation, least privilege and proven cloud-native patterns.

AI agents run untrusted code. Here’s how to secure them using isolation, least privilege and proven cloud-native patterns.


Kubernetes 1.36 launches April 22, 2026, marking a major shift in networking as Ingress-Nginx retires in favor of the more scalable Gateway API. Key updates include bolstered Linux User Namespaces for better isolation and Dynamic Resource Allocation (DRA) enhancements for hardware maintenance. Additionally, manifest-based admission control and OCI artifact mounting aim to boost cluster security.





RapidFort and Nutanix have partnered to integrate automated supply chain security into the Nutanix Kubernetes Platform (NKP). This collaboration enables enterprises to deploy hardened, near-zero CVE container images at scale, accelerating compliance and reducing the attack surface for demanding AI, generative AI, and cloud-native workloads across hybrid environments.


Sidecarless service mesh architectures like Istio Ambient Mode are reducing complexity and reigniting enterprise adoption in 2026.




PagerDuty has extended the capabilities and reach of its artificial intelligence (AI) agents to enable them to be invoked directly from within the Slack messaging platform. Additionally, the AI SRE Agent that is embedded within the PagerDuty Operations Cloud platform can now also leverage the Model Context Protocol (MCP) and an expanded library of application […] [...]

Anthropic has debuted a new capability for its Claude Code platform, called Channels, that enables a number of key improvements. Notably, it allows users to communicate with the coding assistant through messaging platforms like Telegram and Discord. More important, Channels is continuously available, so it is closer to an active collaborator. Clearly, Anthropic’s release of […] [...]

Cloudsmith this week at the KubeCon + CloudNativeCon Europe conference revealed it has added an ability to enrich packages with threat intelligence that enables DevSecOps teams to better evaluate the risk attached to downloading a software component. Nigel Douglas, head of developer relations for Cloudsmith, said this extension to the managed service it provides for […] [...]