MCP EKS Infrastructure Tools

This document describes the Model Context Protocol (MCP) tools available for EKS cluster auditing and infrastructure analysis.

Overview

The EKS Infrastructure MCP Server provides tools to retrieve and analyze Amazon EKS cluster information for auditing and reporting purposes.

Tools

1. getListOfCluster

Description: Retrieves all EKS clusters in the AWS account.

Use Cases:

• Generate audit summaries or reports when cluster information is not available
• List all EKS clusters in the account
• Initial discovery for cluster analysis

Parameters: None

Returns: JSON string containing list of EKS clusters

Example Response:

[
  "cluster-1",
  "cluster-2",
  "production-cluster"
]

2. getEksInfra

Description: Fetches comprehensive EKS cluster infrastructure details including API endpoint, logging configuration, version, encryption settings, and node group information.

Use Cases:

• Generate detailed audit reports for specific clusters
• Analyze cluster infrastructure configuration
• Security and compliance assessments

Parameters:

• cluster_name (string, required): Name of the EKS cluster to analyze

Returns: Structured JSON with complete cluster information

Example Response:

{
  "cluster_name": "my-cluster",
  "status": "ACTIVE",
  "version": "1.28",
  "endpoint": "https://xxx.eks.region.amazonaws.com",
  "roleArn": "arn:aws:iam::account:role/eks-service-role",
  "vpc_config": {
    "subnetIds": ["subnet-xxx", "subnet-yyy"],
    "securityGroupIds": ["sg-xxx"]
  },
  "logging": {
    "clusterLogging": [
      {
        "types": ["api", "audit"],
        "enabled": true
      }
    ]
  },
  "encryptionConfig": [],
  "node_groups": {
    "nodegroup-1": {
      "version": "1.28",
      "status": "ACTIVE",
      "capacityType": "ON_DEMAND",
      "scalingConfig": {
        "minSize": 1,
        "maxSize": 3,
        "desiredSize": 2
      },
      "instanceTypes": ["t3.medium"],
      "subnets": ["subnet-xxx"],
      "remoteAccess": null,
      "amiType": "AL2_x86_64"
    }
  }
}

Setup

Locally

1. Set the AWS_PROFILE environment variable to specify the AWS profile to use
2. Ensure proper AWS credentials and permissions for EKS access
3. Run the MCP server: python mpc_eks_infra.py

UVX

You need to install UV if not already installed

# For Linux and MacOS
curl -LsSf https://astral.sh/uv/install.sh | sh

Config your MCP servers in Claude Desktop, Cursor, ChatGPT Copilot, Github Copilot and other supported AI clients, e.g.

{
  "mcpServers": {
    "aws-eks-infra": {
      "command": "uvx",
      "args": [
        "eks-infra-mcp"
      ],
      "env": {
        "AWS_PROFILE": "<your-aws_profile>"
      }
    }
  }
}

Usage Workflow

1. Discovery: Use getListOfCluster to identify available clusters
2. Analysis: Use getEksInfra with specific cluster names for detailed information
3. Reporting: Process the returned JSON data for audit reports and compliance checks

Error Handling

Both tools return error information in JSON format when exceptions occur:

{
  "error": "Error description"
}