Amazon VPC Lattice MCP Server

A Model Context Protocol (MCP) server for source listing, providing tools for accessing and managing AWS VPC Lattice resources and related documentation.

Features

The server provides five main tools:

1. list_sources: Lists all available sources with their URLs and sample prompts
2. get_source_prompts: Gets sample prompts for a specific source
3. list_amazon_vpc_lattice_prompts: Lists all available prompt templates
4. get_amazon_vpc_lattice_prompts: Gets details of a specific prompt template
5. vpc_lattice_cli: Execute AWS CLI VPC Lattice commands for managing VPC Lattice resources

Installation

This project is built with TypeScript and uses ES modules. Note that installing github-mcp-server is also strongly recommended to assist with development prompts.

1. Clone the repository:

git clone https://github.com/awslabs/amazon-vpc-lattice-mcp-server.git
cd amazon-vpc-lattice-mcp-server

2. Install dependencies:

npm install

3. Build the server:

npm run build

The build script will compile the TypeScript code and set the appropriate executable permissions.

Configuration

Add the server to your MCP settings file (located at ~/Library/Application Support/Code/User/globalStorage/asbx.amzn-cline/settings/cline_mcp_settings.json):

{
  "mcpServers": {
    "amazon-vpc-lattice": {
      "command": "node",
      "args": ["/path/to/amazon-vpc-lattice-mcp-server/build/index.js"],
      "disabled": false,
      "autoApprove": [],
      "env": {}
    }
  }
}

Usage

Once configured, you can use the MCP tools in your conversations. Note that you should use list_amazon_vpc_lattice_prompts to discover available prompts as these are not automatically discoverable like tools.

List Sources

use_mcp_tool({
  server_name: "amazon-vpc-lattice",
  tool_name: "list_sources",
  arguments: {}
})

Get Source Prompts

use_mcp_tool({
  server_name: "amazon-vpc-lattice",
  tool_name: "get_source_prompts",
  arguments: {
    source_name: "AWS Documentation"
  }
})

List Amazon VPC Lattice Prompts

use_mcp_tool({
  server_name: "amazon-vpc-lattice",
  tool_name: "list_amazon_vpc_lattice_prompts",
  arguments: {}
})

Get Amazon VPC Lattice Prompt Details

use_mcp_tool({
  server_name: "amazon-vpc-lattice",
  tool_name: "get_amazon_vpc_lattice_prompts",
  arguments: {
    prompt_name: "setup_eks_controller"
  }
})

VPC Lattice CLI

The vpc_lattice_cli tool provides a programmatic interface to AWS VPC Lattice operations through the AWS CLI.

Features

• Supports all major VPC Lattice CLI operations
• Accepts command arguments as JavaScript objects
• Automatically converts camelCase parameters to CLI-style kebab-case
• Handles boolean flags, arrays, and complex values
• Supports AWS profiles and region configuration
• Returns parsed JSON responses

Available Commands

• Service Network: create-service-network, delete-service-network, get-service-network, list-service-networks, update-service-network
• Service: create-service, delete-service, get-service, list-services, update-service
• Listener: create-listener, delete-listener, get-listener, list-listeners, update-listener
• Rule: create-rule, delete-rule, get-rule, list-rules, update-rule
• Target Group: create-target-group, delete-target-group, get-target-group, list-target-groups, update-target-group
• Target Management: register-targets, deregister-targets, list-targets
• Resource Tags: list-tags-for-resource, tag-resource, untag-resource

Examples

List service networks:

use_mcp_tool({
  server_name: "amazon-vpc-lattice",
  tool_name: "vpc_lattice_cli",
  arguments: {
    command: "list-service-networks",
    region: "us-west-2"
  }
})

Create a service network:

use_mcp_tool({
  server_name: "amazon-vpc-lattice",
  tool_name: "vpc_lattice_cli",
  arguments: {
    command: "create-service-network",
    args: {
      name: "my-network",
      authType: "NONE"
    }
  }
})

Create a service with tags:

use_mcp_tool({
  server_name: "amazon-vpc-lattice",
  tool_name: "vpc_lattice_cli",
  arguments: {
    command: "create-service",
    args: {
      name: "my-service",
      serviceNetworkIdentifier: "sn-12345",
      tags: [
        { key: "Environment", value: "Production" }
      ]
    }
  }
})

Create a target group:

use_mcp_tool({
  server_name: "amazon-vpc-lattice",
  tool_name: "vpc_lattice_cli",
  arguments: {
    command: "create-target-group",
    args: {
      name: "my-target-group",
      type: "INSTANCE",
      config: {
        port: 80,
        protocol: "HTTP",
        healthCheck: {
          enabled: true,
          protocol: "HTTP",
          path: "/health"
        }
      }
    }
  }
})

Available Sources

The server includes these sources:

1. AWS Documentation (docs.aws.amazon.com)

   • Key features queries
   • Configuration guidance
   • Best practices

2. AWS Gateway API Controller for VPC Lattice (aws/aws-application-networking-k8s)

   • Feature support queries
   • Issue tracking

3. Kubernetes Gateway API (gateway-api.sigs.k8s.io)

   • Error resolution
   • Best practices guidance

Development

Project Structure

The project is organized as follows:

• src/index.ts: Main server setup and initialization
• src/tools.ts: Tool definitions and handlers
• src/data/: Data files
  • prompts.ts: Prompt templates and parameters
  • sources.ts: Source definitions and their prompts
• package.json: Project configuration and dependencies
• tsconfig.json: TypeScript configuration
• .gitignore: Git ignore rules
• build/: Compiled JavaScript output

Adding New Sources

To add new sources, modify the sources array in src/data/sources.ts:

export const sources = [
  {
    name: 'Your Source',
    url: 'https://your-source-url.com',
    prompts: [
      'Sample prompt 1 {placeholder}',
      'Sample prompt 2 {placeholder}'
    ]
  }
  // ... existing sources
];

Adding New Prompts

To add new prompt templates, modify the prompts array in src/data/prompts.ts:

export const prompts = [
  {
    name: 'Your Prompt Template',
    description: 'Description of what the prompt does',
    template: 'Your prompt template with {parameter} placeholders',
    parameters: ['parameter']
  }
  // ... existing prompts
];

Scripts

• npm run build: Build the server and set executable permissions
• npm run watch: Watch mode for development
• npm test: Run tests (currently not implemented)