+ Kubernetes × your AI agent Debug a failing pod
If you ask
“The api pod in namespace prod is crash-looping — get the last 100 lines of logs and describe why”
Then
Investigate why a pod is failing by pulling logs and events.
Works in Claude, Codex, Cursor & any MCP agent
How this prompt works
-
Turn on the Kubernetes connector in Claude, ChatGPT or Cursor — once.
-
Paste the prompt above. No code, no menus.
-
Investigate why a pod is failing by pulling logs and events.
Level up: install the Kubernetes skill
A prompt you paste once. A skill your agent keeps — playbooks and defaults it loads every session. Community pack.
npx skills add jeffallan/claude-skills --skill kubernetes-specialist FAQ
How do I debug a failing pod with Kubernetes?
Add the Kubernetes MCP connector once, then paste this prompt into Claude or ChatGPT: "The api pod in namespace prod is crash-looping — get the last 100 lines of logs and describe why". Your AI runs it in Kubernetes.
Do I need to code?
No. You install the connector once (a short config or one click), then ask in plain words. The prompt on this page works as-is.
Which AI assistants work?
Claude (Desktop, Code), Cursor, ChatGPT and other MCP clients. The same prompt works across them.
Set up Kubernetes in Claude
Copy-paste config, setup steps and more ready-to-run prompts on the connector page.
Open Kubernetes setupMore prompts