k8s-debug

To diagnose a Kubernetes pod stuck in CrashLoopBackOff, you need to gather pod status, recent events, container logs (including previous instances), and node context to identify the root cause. This skill automates that collection process.

The best way to collect Kubernetes pod logs and events automatically is using a deterministic, script-driven workflow that retrieves the full pod description, YAML, recent events, and container logs in a reproducible sequence for root-cause analysis.

Yes, you can debug Kubernetes pods across any namespace using kubectl if you have a valid kubeconfig and read permissions for pods, events, and nodes. This skill uses that access to run its diagnostic workflows.

You need minimal privilege RBAC permissions to run Kubernetes pod diagnostics. Specifically, you require read permissions for pods, events, and nodes, along with standard kubectl access and a valid kubeconfig.

To troubleshoot Kubernetes pod DNS and networking issues, you need to gather comprehensive diagnostics including pod status, events, and node context. This skill automates that data collection to support root-cause analysis.

This approach to Kubernetes debugging starts with read-only exploration to ensure safe guidance for disruptive actions. It collects diagnostics predictably and only proceeds with disruptive actions after receiving explicit confirmation.