auto-om

auto-om is a Python-based MCP server that provides a large suite of Linux automation tools accessible over the MCP protocol. It exposes 74+ tools organized into categories such as File & Directory Management, Text Processing & Analysis, System Monitoring, Process Management, Network Operations, Compression & Archive, and Package Management. The server runs in either stdio or HTTP mode and communicates via MCP-compatible clients, enabling you to remotely manage Linux hosts through SSH connections. To get started, run the server (for example in HTTP mode) and connect with an MCP client to discover and invoke the available tools. Typical interactions involve requesting file operations, monitoring system metrics, managing processes, performing network diagnostics, handling archives, and managing packages on configured Linux hosts. The server is designed for non-root operations with a focus on safe SSH-based execution and dry-run capabilities to preview destructive actions.

Prerequisites:

• Python 3.11 or higher
• pip and optionally pipenv
• Linux servers with SSH access (for remote operations)

Setup from Source:

# Clone the repository
git clone https://github.com/atoncooper/auto-om.git
cd auto-om

# Install dependencies with pipenv (recommended)
pipenv install

# Activate virtual environment
pipenv shell

# Or install dependencies with pip directly
pip install -r requirements.txt

Using Docker (alternative):

# Build the Docker image
docker build -t auto-om:latest .

# Run with Docker (example)
docker run -p 8000:8000 -v $(pwd)/application.yaml:/app/application.yaml auto-om:latest

Usage:

# Start in stdio mode (default)
python main.py

# Start in HTTP mode
python main.py --mode http

# Run with a custom config file
python main.py --config /path/to/application.yaml

Configuration: Create an application.yaml as shown in the README to configure transport mode, HTTP settings, and SSH pool with your servers.

Tips:

• Security: Use environment variables for sensitive data (e.g., SSH passwords) and reference them in application.yaml as ${ENV_VAR}.
• The server uses an SSH connection pool and enforces non-root operations where possible.
• If integrating with MCP clients (e.g., Claude Code, Cursor, Trae), ensure the client is configured to connect to http://<server>:8000/sse or the appropriate MCP endpoint.
• For Docker deployments, you can mount a custom application.yaml to override defaults.

Common issues:

• Module not found errors: ensure PYTHONPATH includes the src directory or install in editable mode.
• SSH authentication failures: verify network connectivity, credentials, and that the target servers allow SSH access.
• Resource limits: adjust max_connections and timeouts in ssh_pool if managing many hosts.

Environment variables and configuration options:

• SSH_PASSWORD or other credentials via environment variables helps avoid embedding secrets in YAML.
• Transport mode can be http or stdio; http mode exposes an HTTP endpoint for MCP clients.