Dockhand

An MCP server that gives an LLM (Claude, etc.) direct control over your local Docker daemon — run and manage containers, images, volumes, networks, and Compose stacks through natural language.

Built with FastMCP and the Docker SDK for Python.

Requirements

• Python >= 3.14

• Docker installed and running locally (Docker Desktop, Docker Engine, etc.)

• Docker Compose v2 CLI (docker compose) available on PATH — required for the stack/compose tools

• uv for dependency management

Related MCP server: Docker MCP Server

Installation

git clone <this-repo>
cd dockhand
uv sync

This creates a .venv and installs the dependencies pinned in uv.lock.

Running

uv run main.py

The server communicates over stdio, so it's meant to be launched by an MCP client rather than run standalone in a terminal.

Connecting to Claude Desktop / Claude Code

Add an entry to your MCP client's configuration (e.g. claude_desktop_config.json):

{
  "mcpServers": {
    "dockhand": {
      "command": "uv",
      "args": ["--directory", "/absolute/path/to/dockhand", "run", "main.py"]
    }
  }
}

Restart the client and the tools below become available to the model.

Available tools

Containers (containers.py)

Images (images.py)

Volumes & networks (volumes.py)

Compose stacks (stacks.py)

Project structure

main.py            # Entrypoint: registers tool modules and starts the MCP server
server.py           # Shared FastMCP server instance
docker_client.py    # Lazy singleton Docker SDK client
compose_client.py   # Thin wrapper around the `docker compose` CLI
containers.py       # Container lifecycle & inspection tools
images.py           # Image build/push/list/delete tools
volumes.py          # Volume & network tools
stacks.py           # Compose stack tools

Safety notes

This server gives the model real, unsandboxed control over your Docker daemon:

• delete_container and delete_image force-remove resources without confirmation.

• remove_stack deletes volumes (-v), which is destructive and irreversible for stateful data.

• container_exec runs arbitrary shell commands inside a container.

• push_image pushes to whatever registry you point it at, using your existing local Docker credentials.

Only connect this server to clients/agents you trust, and be deliberate about which containers and stacks you let it touch.

License

No license specified.