Inspektor Gadget MCP Server

AI-powered debugging and inspection for Kubernetes clusters using Inspektor Gadget.

https://github.com/user-attachments/assets/86367982-c0aa-455c-ac9e-ca43348899df

Features

• AI-powered interface for Kubernetes troubleshooting and monitoring

• One-click Inspektor Gadget deployment and removal

• Intelligent output summarization and analysis

• Automatic gadget discovery from Artifact Hub

Related MCP server: Azure MCP Server

Quick Start

1. Ensure you have Docker and a valid kubeconfig file

2. Configure the MCP server in VS Code (see INSTALL.md)

3. Start using AI commands in VS Code Copilot Chat

4. Try: "Show me DNS traffic" or "Deploy Inspektor Gadget"

5. Head to examples for detailed examples.

Installation

You can use the following commands to quickly configure the Inspektor Gadget MCP server using either Docker or a binary in your VS Code settings.

Kubernetes

To use IG MCP Server you will need to start by deploying Inspektor Gadget using one of the following commands:

IG_VERSION=$(curl -s https://api.github.com/repos/inspektor-gadget/inspektor-gadget/releases/latest | jq -r '.tag_name' | sed 's/^v//')
helm install gadget --namespace=gadget --create-namespace oci://ghcr.io/inspektor-gadget/inspektor-gadget/charts/gadget --version=$IG_VERSION

or

IG_VERSION=$(curl -s https://api.github.com/repos/inspektor-gadget/inspektor-gadget/releases/latest | jq -r '.tag_name')
kubectl apply -f https://github.com/inspektor-gadget/inspektor-gadget/releases/download/${IG_VERSION}/inspektor-gadget-${IG_VERSION}.yaml

Then you can deploy the IG MCP server to a Kubernetes cluster using the following command:

kubectl apply -f https://raw.githubusercontent.com/inspektor-gadget/ig-mcp-server/main/manifests/ig-mcp-server-all.yaml

Then use the following command to port-forward the server to your local machine:

kubectl port-forward svc/ig-mcp-server 8080:8080 -n gadget

Then you can use http://localhost:8080/mcp as a remote MCP url (transport: http) in your agent.

Security Note: The IG MCP server does not currently implement built-in authentication. For production deployments, it is strongly recommended to place the server behind a reverse proxy (such as nginx, Traefik, or an API gateway) with proper authentication and TLS termination. Additionally, ensure that the service is not directly exposed to the internet and use network policies to restrict access to trusted sources only.

Docker

Binary

You can head to the Releases page and download the latest binary for your platform:

After downloading, you can run the following command to add it to your VS Code MCP configuration.

Available Tools

Inspektor Gadget Lifecycle

• ig_deploy: Manage deployment of Inspektor Gadget on target Kubernetes cluster

Gadget Lifecycle

• ig_gadgets: Manage the lifecycle of running gadgets (stop/list/get-results)

Dynamic Tools

Each gadget is registered as its own MCP tool (e.g., gadget_trace_dns, gadget_trace_tcp, etc.) and supports running gadgets in foreground mode, which is useful for debugging/development and also in background mode for observability.

Also, You can control which gadgets are available by configuring the MCP server with the -gadget-discoverer or -gadget-images options, allowing you to limit the tools to only those you need.

Gadget Discovery

Gadget discovery allows controlling which gadgets are available for use. You can choose between two methods:

• Automatic: Uses Artifact Hub (-gadget-discoverer=artifacthub)

• Manual: Specify gadgets directly (-gadget-images=trace_dns:latest)

See INSTALL.md for configuration options.

Security Notes

• Requires read-only access to your kubeconfig file

• Needs network access for Artifact Hub discovery

• See security guide for setting up server with minimal permissions

Resources

• 📖 Documentation

• 🔍 Examples

• 🐛 Issues

• 💬 Slack

• 🌐 Website

• 📋 Troubleshooting

• 🔒 Security Guide

Related Projects

• Inspektor Gadget - Kubernetes debugging tool

• MCP Specification - Model Context Protocol

License

Apache License 2.0 - see LICENSE for details.