kubectl-mcp-server and ig-mcp-server
These tools are competitors, as both implement a Model Context Protocol (MCP) server for Kubernetes, with tool B offering AI integration for debugging.
About kubectl-mcp-server
rohitg00/kubectl-mcp-server
A Model Context Protocol (MCP) server for Kubernetes. Install: npx kubectl-mcp-server or pip install kubectl-mcp-server
Exposes 253 Kubernetes tools through MCP's transport layer, enabling AI assistants to perform diagnostics, deployments, and cost optimization via natural language rather than imperative commands. Supports multi-cluster management, interactive HTML dashboards, and non-destructive mode with OAuth 2.1 authentication for enterprise deployments. Integrates with Claude, Cursor, Windsurf, Copilot and 15+ AI platforms through standardized MCP protocol.
About ig-mcp-server
inspektor-gadget/ig-mcp-server
Debug your Container and Kubernetes workloads with an AI interface
Exposes Inspektor Gadget's eBPF-powered kernel observability (DNS traces, TCP connections, syscalls, file activity) as dynamically-registered MCP tools, enabling LLMs to autonomously select relevant gadgets and correlate low-level telemetry across multiple data sources. Supports foreground/background execution modes, read-only production safety, and flexible deployment via stdio, Docker, or Kubernetes HTTP transport. Transforms raw kernel events into structured JSON enriched with Kubernetes metadata (pod, namespace, container), enabling AI-driven root cause analysis grounded in real evidence rather than speculation.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work