Walkthrough: How we containerized a Claw agent for our Kubernetes cluster.

I saw the blog post about containerizing a Claw agent for Kubernetes. We're looking at similar agent-based tools for security scanning, so this was really timely.

The post mentions they used a multi-stage Docker build to keep the image small. I get why that's important, but I'm still fuzzy on the networking part. How do you handle the agent calling home from inside the cluster? Did they need a special egress rule or a sidecar? Also, how do you manage the config secrets for the agent across different environments? Is it just a Kubernetes Secret, or something more complex?