MCP Ansible Automation Platform Server

A Very basic and simple Model Context Protocol (MCP) server that provides integration with Ansible Automation Platform (AAP) via API. This server allows Large Language Models to interact with AAP to extract available templates and launch Ansible job templates with custom parameters. This is just a demo, and not a full MCP Server capabiliy, feel free to use this code as you want.

Features

• Template Discovery: Extract available job templates from configured AAP projects

• Job Launching: Launch Ansible job templates with custom extra variables and parameters

• Job Monitoring: Check job status and retrieve job outputs/logs

• Connection Testing: Verify AAP connectivity and authentication

• Error Handling: Robust error handling with retry logic and detailed error messages

Related MCP server: AAP MCP Server

Prerequisites

• Python 3.8+

• Access to an Ansible Automation Platform instance

• AAP API access token with appropriate permissions

AAP Token Generation

Installation

1. Clone this repository

2. Install dependencies:

   uv pip install -r requirements.txt

3. Configure environment variables:

   cp env_example .env
   # Edit .env with your AAP configuration

Configuration

Create a .env file in the project root with the following variables:

# Ansible Automation Platform Configuration
AAP_URL=https://your-aap-instance.com
AAP_TOKEN=your-access-token-here
AAP_PROJECT_ID=your-project-id
AAP_VERIFY_SSL=True

# Optional settings
AAP_TIMEOUT=30
AAP_MAX_RETRIES=3

Configuration Parameters

• AAP_URL: Base URL of your AAP instance (e.g., https://controller.example.com)

• AAP_TOKEN: Bearer token for API authentication

• AAP_PROJECT_ID: Default project ID to extract templates from

• AAP_VERIFY_SSL: Whether to verify SSL certificates (True/False)

• AAP_TIMEOUT: Request timeout in seconds (default: 30)

• AAP_MAX_RETRIES: Number of retry attempts for failed requests (default: 3)

Usage

Testing the Server

Before integrating with Claude Desktop, test your server setup:

uv run python tools/test_mcp_server.py

This will validate:

• Environment configuration

• AAP connection

• Template extraction

• MCP server functionality

Running the Server

Start the MCP server:

uv run python server.py

The server will start and listen for MCP protocol messages on stdin/stdout.

Claude Desktop Integration

Step 1: Configure Claude Desktop

1. Locate your Claude Desktop configuration file:

   • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

   • Windows: %APPDATA%\Claude\claude_desktop_config.json

   • Linux: ~/.config/claude/claude_desktop_config.json

2. Edit the configuration file and add the MCP server:

{
  "mcpServers": {
    "ansible-aap": {
      "command": "uv",
      "args": ["run", "python", "server.py"],
      "cwd": "<PATH TO THE PROJECT>",
      "env": {
        "AAP_URL": "https://your-aap-instance.com",
        "AAP_TOKEN": "your-access-token-here",
        "AAP_PROJECT_ID": "your-project-id",
        "AAP_VERIFY_SSL": "True",
        "AAP_TIMEOUT": "30",
        "AAP_MAX_RETRIES": "3"
      }
    }
  }
}

Important: Replace the placeholder values with your actual AAP configuration:

• cwd: Use the absolute path to your MCP server directory

• AAP_URL: Your AAP instance URL

• AAP_TOKEN: Your API access token

• AAP_PROJECT_ID: Your project ID

Step 2: Restart Claude Desktop

After saving the configuration, restart Claude Desktop completely.

Step 3: Test the Integration

Once restarted, you can test the integration by asking Claude questions like:

• "What Ansible templates are available?"

• "Show me the job templates from my AAP project"

• "Show nodes with with low disk space"

• "Check the status of job 123"

Step 4: Verify Connection

You should see Claude respond with information about your Ansible templates and be able to launch jobs. If there are issues, check:

1. Claude Desktop logs for error messages

2. Your AAP credentials and permissions

3. Network connectivity to your AAP instance

Alternative: Using Environment File

Instead of embedding credentials in the Claude Desktop config, you can create a .env file:

{
  "mcpServers": {
    "ansible-aap": {
      "command": "uv",
      "args": ["run", "python", "server.py"],
      "cwd": "<PATH TO THE PROJECT>"
    }
  }
}

Then create a .env file in your project directory with the AAP configuration.

Available Tools

The server provides the following tools:

1. get_job_templates

Extract available job templates from the configured AAP project.

Parameters:

• project_id (optional): Override the configured project ID

Example:

{
  "name": "get_job_templates",
  "arguments": {
    "project_id": "5"
  }
}

2. launch_job_template

Launch an Ansible job template with optional parameters.

Parameters:

• template_id (required): ID of the job template to launch

• extra_vars (optional): Extra variables to pass to the job template

• inventory (optional): Inventory ID to use

• credentials (optional): List of credential IDs to use

Example:

{
  "name": "launch_job_template",
  "arguments": {
    "template_id": 10,
    "extra_vars": {
      "target_host": "web-server-01",
      "service_name": "nginx",
      "restart_service": true
    },
    "inventory": 2
  }
}

3. get_job_status

Get the status and details of a running or completed job.

Parameters:

• job_id (required): ID of the job to check

Example:

{
  "name": "get_job_status",
  "arguments": {
    "job_id": 123
  }
}

4. get_job_output

Get the output/logs of a job.

Parameters:

• job_id (required): ID of the job to get output from

Example:

{
  "name": "get_job_output",
  "arguments": {
    "job_id": 123
  }
}

5. test_aap_connection

Test the connection to Ansible Automation Platform.

Parameters: None

Example:

{
  "name": "test_aap_connection",
  "arguments": {}
}

Integration with LLM Clients

This server can be integrated with various MCP-compatible LLM clients. The server provides a standardized interface for LLMs to:

1. Discover Infrastructure: Query available Ansible templates to understand what automation is available

2. Execute Operations: Launch specific templates with custom parameters based on the conversation context

3. Monitor Progress: Check job status and retrieve outputs to provide feedback to users

Example LLM Interaction Flow

1. Discovery: LLM calls get_job_templates to see what automation is available

2. Planning: Based on user request, LLM identifies the appropriate template and parameters

3. Execution: LLM calls launch_job_template with the required parameters

4. Monitoring: LLM uses get_job_status and get_job_output to track progress and provide updates

Error Handling

The server includes comprehensive error handling:

• Connection Errors: Automatic retry with exponential backoff

• Authentication Errors: Clear error messages for token/permission issues

• Validation Errors: Parameter validation with helpful error messages

• Rate Limiting: Respectful API usage with configurable timeouts

Security Considerations

• Store AAP tokens securely in environment variables

• Use HTTPS for AAP connections

• Validate SSL certificates in production (AAP_VERIFY_SSL=True)

• Limit template access through AAP project configuration

• Monitor job launches and outputs for security compliance