Kubernetes

This MCP server provides comprehensive Kubernetes cluster management capabilities, enabling AI assistants to list, get, apply, delete, and execute commands on Kubernetes resources across both namespaced and clustered scopes. Built by StacklokLabs using Go with the Kubernetes client-go library and mark3labs/mcp-go framework, it features rate limiting with configurable per-tool limits, timeout middleware to prevent context cancellation errors, periodic kubeconfig refresh for long-running deployments, and special handling for pod operations including log retrieval and command execution with configurable timeouts. The implementation offers both read-only and read-write modes, optional resource serving to reduce LLM context size for large clusters, SSE transport support, and tools for applying YAML manifests, retrieving pod logs with filtering options, executing commands in containers, and accessing subresources like status and scale, making it valuable for DevOps automation, cluster monitoring and troubleshooting, CI/CD pipeline integration, and any workflow requiring programmatic Kubernetes cluster management through natural language interfaces.

Source

Repository: https://github.com/stackloklabs/mkp