Kubernetes Skills
Use this skill for Kubernetes cluster management, workload operations, and troubleshooting.
Workflow
- Use
kubectlfor standard operations. - For follow-up repo troubleshooting, read the latest relevant entries in
tmp/agent-handoff.mdbefore starting new cluster triage so prior findings and failed paths are not repeated. - If available, use Kubernetes MCP tools for deeper cluster insight.
- Read upstream-current.md when the task touches version-sensitive Kubernetes, HPA behavior, or AgentGateway-on-Kubernetes.
- Use
https://agentgateway.dev/docs/kubernetes/latest/for AgentGateway-on-Kubernetes setup, configuration, and runtime troubleshooting. - Use official Kubernetes docs (
https://kubernetes.io/docs) andkubectlhelp for core API behavior. - For troubleshooting, inspect events, pod status,
describe, and logs before suggesting architectural changes.
Tooling
- Use
kubectlvia terminal. - Use Kubernetes MCP server (if configured).
- Prioritize documentation in this order: AgentGateway Kubernetes docs for gateway-specific tasks, then official Kubernetes docs.
Source: Shkirmantsev/AgenticNativePlatform — distributed by TomeVault.