mcp · kubeshark — Claude Code MCP server: what it does, how to install
This MCP server gives AI agents access to Kubernetes network traffic. Use it to inspect API calls, investigate incidents, and trace failures across services.
Open on GitHubWhat mcp does
mcp connects an AI agent to cluster-wide Kubernetes network data. It can query traffic, inspect API calls, and support root cause analysis through natural-language requests. The listed protocol coverage includes HTTP, gRPC, Redis, Kafka, and DNS.
Kubeshark indexes traffic at the kernel level with eBPF and combines Kubernetes identity, API context, and network attributes in queries. The stack also provides traffic snapshots and PCAP export, including decrypted TLS traffic.
The kubeshark CLI must be installed first because the MCP command runs kubeshark mcp. The package also includes AI skills for retrospective root cause analysis and KFL traffic filters, available as a Claude Code plugin.
What it helps with
- trace a failed request across Kubernetes services
- inspect API calls behind an incident
- filter traffic by workload and network attributes
- export a PCAP for Wireshark analysis
- compare service errors and TCP retransmissions
Example prompts for the agent
- Why did the checkout request fail at 2:15 PM? Investigate it using the cluster traffic.
- Trace request
abc123through every service and show where it failed. - Compare service errors and show TCP retransmission rates between cluster nodes.
How to install mcp
Install
The MCP registry has no launch command for it: see the repository README
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about mcp
How do I install mcp in Claude Code?
Install the CLI with brew install kubeshark, then add the server to Claude Code with claude mcp add kubeshark -- kubeshark mcp.
What does mcp do?
It exposes Kubernetes network data to an AI agent through MCP. The agent can inspect traffic, analyze API calls, and investigate failure causes.
Which protocols does mcp support?
The listed protocols are HTTP, gRPC, Redis, Kafka, and DNS. Queries can combine Kubernetes identity, API context, and network attributes.
