Security | Threat Detection | Cyberattacks | DevSecOps | Compliance

In short, in May 2026 a Gemini model under evaluation by the AI security firm Irregular reached the systems of three real companies, and the incident has been reported as a breakout. By Google’s own account it was nothing of the kind. The test environment had internet access it was not meant to have, the fictional target shared its name with a real company, and the model found public information online, guessed one password and found two more in public code repositories.

In short, the EU Cyber Resilience Act (CRA) puts binding cybersecurity requirements on any software sold as a product in the EU, and it does not carve out AI agents. Since 11th September 2026, any company that sells software with digital elements into the EU has 24 hours from learning that a vulnerability is being exploited to file an early warning with a national CSIRT and ENISA, 72 hours to describe it, and 14 days to report what it did about it.

A library of Calico tools and skills — delivered through the Calico MCP Server.

One of the blockers to moving VMs off vSphere and onto Kubernetes is losing NSX and the protection it provides. Security teams that have spent years building out distributed firewall policy look at Kubernetes and are, quite understandably, alarmed by the flat network and the fact that any workload can reach any other by default. How will they enforce east-west traffic controls? Will they be able to replicate NSX distributed firewall rules with the same granularity?

In short, the MCP 2026-07-28 release candidate is getting attention for going stateless. The quieter story is a package of six SEPs that harden the protocol’s OAuth layer: issuer validation, credential binding, client type declaration, and cleanups around refresh tokens, scopes, and discovery. All six are worth shipping, and all six fix real failure modes. But they harden how a client authenticates to a server, and that was never the whole problem.

Most attempts to control AI agents work at the model layer (alignment, system prompts) or the application layer (guardrail libraries, output filters). Both share a flaw: the thing being secured is also the thing doing the securing. A sufficiently confused or sufficiently compromised agent can talk its way past its own instructions. OpenShell takes a different position, and it is the right one. Put the controls in the environment, where the agent cannot negotiate with them.

As Kubernetes clusters scale from a few development sandboxes to massive, multi-tenant production environments, platform teams often find themselves facing a configuration management crisis. A small number of microservices suddenly demand hundreds of individual Kubernetes NetworkPolicy objects. Managing them becomes operationally expensive, auditing them is difficult, and a single developer misconfiguration can easily drop critical production traffic or open a massive security hole.

As part of our work at Tigera building products that create secure runtime environments for enterprise agents at scale in the real world, one small part of this puzzle I think about a lot is policy, and runtime enforcement of policy, and how to create a comprehensive secure runtime, configured from one place.

Practically no one runs a single Kubernetes cluster in production these days. Maybe that’s how it started but data sovereignty requirements, acquisitions, AI initiatives and the need for edge servers, among other considerations, have pulled most enterprises into multi-cluster territory whether they planned for it or not.

Note: The core architecture for this pattern was introduced by Isaac Hawley from Tigera. If you are building an AI agent that relies on tool calling, complex routing, or the Model Context Protocol (MCP), you’re not just building a chatbot anymore. You are building an autonomous system with access to your internal APIs. With that power comes a massive security and governance headache, and AI agent security testing is where most teams hit a wall.