MCP.so
Sign In

Kubernetes MCP Server

@ehiaig

About Kubernetes MCP Server

An MCP server for debugging and managing Kubernetes using Natural Language

Basic information

Category

Cloud & Infrastructure

Runtime

python

Transports

stdio

Publisher

ehiaig

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "kubernetes-mcp-server": {
      "command": "/path/to/your/.local/bin/uv",
      "args": [
        "--directory",
        "/path/to/this/repo/kubernetes-mcp-server/src/k8s_manager",
        "run",
        "-m",
        "k8s_manager"
      ]
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Kubernetes MCP Server?

An MCP server that lets you debug and manage Kubernetes configuration files and clusters using natural language. It can analyze a single YAML config, a folder of YAML files, and connect to clusters to manage pods (list, get, logs, delete). It is designed for developers and platform engineers who want to interact with Kubernetes through conversational AI.

How to use Kubernetes MCP Server?

Install the server via uv pip install git+https://github.com/ehiaig/kubernetes-mcp-server or by cloning the repo. Then add a corresponding entry to your Claude Desktop configuration file (claude_desktop_config.json) with the uv command and appropriate arguments. After configuring, you can send natural language queries to the server through Claude.

Key features of Kubernetes MCP Server

  • Debug a single Kubernetes config YAML file
  • Debug an entire Kubernetes config folder
  • Pod management: list, get, logs, delete
  • Connect to and manage Kubernetes clusters via natural language

Use cases of Kubernetes MCP Server

  • Quickly identify misconfigurations in a deployment YAML file
  • Review an entire directory of Kubernetes manifests for errors
  • Troubleshoot a failing pod by fetching logs and describing resources
  • Ask for a summary of running pods in a cluster without using kubectl

FAQ from Kubernetes MCP Server

How do I install the server?

Install using uv pip install from the repository or clone the repo locally. Then add the server definition to your Claude Desktop JSON configuration file.

Can I manage deployments or services?

Only pod management is implemented at this time. Deployment, service, and namespace management are planned but not yet available.

Is it safe to use with sensitive data?

No. Do not configure clusters that contain secrets, passwords, or other sensitive data. Any sensitive information exchanged with the LLM may be compromised unless the LLM runs entirely on your local machine.

What runtime dependencies are required?

Python, uv, and a Claude Desktop client. The server is run via uv from the cloned repository or installed package.

How do I configure the server in Claude Desktop?

Edit claude_desktop_config.json (located at ~/Library/Application Support/Claude/ on macOS or %APPDATA%/Claude/ on Windows) and add an entry with the uv command pointing to the server's module.

Comments

More Cloud & Infrastructure MCP servers