kubernetes · bobmatnyc — Claude Code skill: deploy and debug in a cluster
A Kubernetes operations playbook for your agent: safe rollouts, health probes, resource sizing and fast kubectl triage.
npx skills add bobmatnyc/claude-mpm-skills --skill kubernetes -a claude-codeWhat kubernetes does
The skill hands your agent a set of rules for running Kubernetes workloads. It covers picking the right workload and controller, wiring readinessProbe and startupProbe so a rollout doesn't drop traffic, and setting requests and limits — skip those and neighbouring pods start fighting over the node. Security gets its own section: run as non-root, grant least privilege.
Quick start is built on kubectl. The agent inspects the pod, pulls events from the namespace, then reads logs from the previous container — the usual order when something won't come up. Four references sit behind that: references/core-objects.md, references/rollouts-and-probes.md, references/debugging-runbook.md and references/security-hardening.md. You need kubectl installed and a reachable cluster, or none of the commands run.
There is no slash command here. The frontmatter sets user-invocable: false, so the skill loads on its own when the agent touches a deploy or a diagnosis. Worth knowing if you were expecting to trigger it by hand.
What it helps with
- find out why a pod never reaches Running
- set requests and limits so pods stop evicting each other
- add probes before shipping a new version
- read namespace events in chronological order
- check whether a pod runs as root
Example prompts for the agent
- write a deployment manifest with readinessProbe and startupProbe
- the pod won't start, check describe output and namespace events
- review requests and limits on this deployment
How to install kubernetes
1Skills CLI (Vercel)
npx skills add bobmatnyc/claude-mpm-skills --skill kubernetes -a claude-code2Manually: put the files in place
~/.claude/skills/kubernetes/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/kubernetes/
1Skills CLI (Vercel)
npx skills add bobmatnyc/claude-mpm-skills --skill kubernetes -a codex2Manually: put the files in place
.agents/skills/kubernetes/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add bobmatnyc/claude-mpm-skills --skill kubernetes -a cursor2Manually: put the files in place
.cursor/skills/kubernetes/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install bobmatnyc/claude-mpm-skills kubernetes2Manually: put the files in place
.github/skills/kubernetes/SKILL.md1Skills CLI (Vercel)
npx skills add bobmatnyc/claude-mpm-skills --skill kubernetes -a gemini-cli2Platform CLI
gemini skills install https://github.com/bobmatnyc/claude-mpm-skills --path universal/infrastructure/kubernetesManually: put the files in place
.devin/skills/kubernetes/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add bobmatnyc/claude-mpm-skills --skill kubernetes -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about kubernetes
How do I install the kubernetes skill in Claude?
Run npx skills add bobmatnyc/claude-mpm-skills --skill kubernetes -a claude-code. The skill file then lives at ~/.claude/skills/kubernetes/SKILL.md.
What does the kubernetes skill do?
It guides the agent through Kubernetes operations: choosing objects, configuring probes, sizing resources, rolling out safely and triaging failures with kubectl.
Do I have to invoke the skill manually?
No. It sets user-invocable: false, so it loads by itself whenever the agent works on a deploy or cluster diagnosis.
Source text
SKILL.md · MIT
---
name: kubernetes
description: "Kubernetes operations playbook for deploying services: core objects, probes, resource sizing, safe rollouts, and fast kubectl debugging"
user-invocable: false
disable-model-invocation: true
version: 1.0.0
category: universal
author: Claude MPM Team
license: MIT
progressive_disclosure:
entry_point:
summary: "Operate Kubernetes workloads with safe rollouts, health probes, resource sizing, and fast kubectl debugging"
when_to_use: "When deploying services to Kubernetes, diagnosing cluster/runtime issues, or hardening workloads for production readiness"
quick_start: "1. Inspect: kubectl get/describe 2. Check events/logs 3. Add probes + requests/limits 4. Roll out safely 5. Validate endpoints"
token_estimate:
entry: 150
full: 9000
context_limit: 900
tags:
- kubernetes
- k8s
- infrastructure
- deployment
- operations
- reliability
requires_tools:
- kubectl
---
# Kubernetes
## Quick Start (kubectl)
```bash
kubectl describe pod/<pod> -n <ns>
kubectl get events -n <ns> --sort-by=.lastTimestamp | tail -n 30
kubectl logs pod/<pod> -n <ns> --previous --tail=200
```
## Production Minimums
- Health: `readinessProbe` and `startupProbe` for safe rollouts
- Resources: set `requests`/`limits` to prevent noisy-neighbor failures
- Security: run as non-root and grant least privilege
## Load Next (References)
- `references/core-objects.md` — choose the right workload/controller and service type
- `references/rollouts-and-probes.md` — probes, rollouts, graceful shutdown, rollback
- `references/debugging-runbook.md` — common failure modes and a fast triage flow
- `references/security-hardening.md` — pod security, RBAC, network policy, supply chain
© Claude MPM Team · MIT · bobmatnyc/claude-mpm-skills
Similar skills
Everything for DevOps and infrastructure →kubernetes by other authors comes first.
