---
title: "kubernetes"
description: "Inspect and diagnose Kubernetes clusters: pods, resources, events, logs, and health, via a trusted read-only Kubernetes MCP server plus workflow skills."
canonical: https://agentpluginsdirectory.com/plugins/kubernetes
last-updated: 2026-09-22
---

# kubernetes
Inspect and diagnose Kubernetes clusters: pods, resources, events, logs, and health, via a trusted read-only Kubernetes MCP server plus workflow skills.
- Slug: kubernetes
- Publisher: HiAI
- Repository: https://github.com/HiAi-gg/agent-plugins
- Manifest: plugins/kubernetes/plugin.json
- Version: 0.0.2
- License: MIT
- Category (editorial): devops-cloud
- Skills: 3 (diagnose-pod, explain-kubernetes-resource, review-cluster)
- MCP servers: 1 (kubernetes)
- Stars: 1
- Repository created: 2026-08-08
- Repository last pushed: 2026-09-03
- Publisher type: Organization
- Listing: https://agentpluginsdirectory.com/plugins/kubernetes
- Schema: https://agent-plugins.org/schemas/1.0.0/plugin.schema.json

## What kubernetes does, in the publisher's words

Inspect and diagnose Kubernetes clusters: pods, resources, events, logs, and health, via a trusted read-only Kubernetes MCP server plus workflow skills.

An Agent Plugin (spec 1.0.0).

A Skills + MCP plugin:

- MCP server (kubernetes): kubernetes-mcp-server@0.0.65 launched with --read-only and --disable-multi-cluster, so only read-annotated tools are exposed.
- Skills: diagnose-pod, review-cluster, explain-kubernetes-resource: read-only workflows for pod debugging, cluster health review, and resource explanation.
- The MCP server gives typed tool access; the skills teach the investigation order (evidence → cause → next step).
- --read-only is enforced by the server, not just by the skills: verified against a disposable cluster (see below).
- No credentials stored: uses the user's existing kubeconfig/current context.
- Server: kubernetes: stdio, npx -y kubernetes-mcp-server@0.0.65 --read-only --disable-multi-cluster (pinned).

From the project README, punctuation lightly normalized. Full text: https://raw.githubusercontent.com/HiAi-gg/agent-plugins/HEAD/plugins/kubernetes/README.md

## Skills

- diagnose-pod: Debug a failing, unhealthy, CrashLoopBackOff, or ImagePullBackOff Kubernetes pod. Use when a pod is not Ready, restarting, stuck in Pending, or when a workload is misbehaving.
- explain-kubernetes-resource: Explain what a Kubernetes resource does, how its spec maps to runtime objects, and how workloads relate (Deployments, Services, Ingress, ConfigMaps, Secrets, PVCs). Use when the user asks what a YAML manifest means, what a resource does, or how objects connect.
- review-cluster: Perform a read-only health review of a Kubernetes cluster or namespace: node status, workload state, resource pressure, events, and obvious misconfigurations. Use when asked for an overall cluster or namespace health assessment.

Descriptions come from the frontmatter of each SKILL.md, punctuation lightly normalized.

## MCP servers

- kubernetes: transport: stdio; command: npx -y kubernetes-mcp-server@0.0.65 --read-only --disable-multi-cluster

Read from the plugin's own mcp.json. Environment variable names only, never values.