Kubernetes Pod Diagnostic Agent
Diagnoses Kubernetes pod failures using kubectl and the Kubernetes API server endpoints. Analyzes CrashLoopBackOff, OOMKilled, and ImagePullBackOff states by querying /api/v1/namespaces/{ns}/pods/{pod}/log and /api/v1/events resources.
npx skills add agentskillexchange/skills --skill kubernetes-pod-diagnostic-agent
The Kubernetes Pod Diagnostic Agent skill provides automated troubleshooting for failing Kubernetes workloads. It interfaces with the Kubernetes API server through kubectl commands and direct REST API calls to /api/v1 and /apis/apps/v1 endpoints to diagnose common pod failure patterns.
What this skill actually does
The skill systematically investigates pod issues by first querying pod status via /api/v1/namespaces/{ns}/pods/{name} to identify the failure state (CrashLoopBackOff, OOMKilled, ImagePullBackOff, Pending, CreateContainerConfigError). It then retrieves container logs using the pod log subresource with previous=true for crashed containers and tailLines for recent output.
Diagnostic capabilities include analyzing resource quotas and LimitRange conflicts via /api/v1/namespaces/{ns}/resourcequotas, checking node capacity and taints through /api/v1/nodes, inspecting PersistentVolumeClaim bindings, verifying ServiceAccount and RBAC permissions via /apis/rbac.authorization.k8s.io, and reviewing NetworkPolicy rules that may block pod communication. The skill generates structured runbook output with root cause analysis, remediation commands, and links to relevant Kubernetes documentation. It also checks for common misconfigurations in Deployment, StatefulSet, and DaemonSet specs.