Model Context Protocol moderate risk

Docker

Docker-mcp manages Docker containers and Compose stacks via Claude AI, offering container lifecycle management and log retrieval.

Connections & Capabilities

Connects To

GitHubDocker

Capabilities

readwriteexec

Quickstart

Install

npx @smithery/cli install docker-mcp --client claude

Config

{
  "mcpServers": {
    "docker-mcp": {
      "command": "uvx",
      "args": [
        "docker-mcp"
      ]
    }
  }
}

Exposed MCP Tools (4)

moderate

create-container

Creates a Docker container with specified image, name, ports, and environment variables.

Can create containers with potentially insecure configurations.

high

deploy-compose

Deploys a Docker Compose stack using a provided YAML file.

Compose files can contain arbitrary configurations, posing significant security risks.

safe

get-logs

Retrieves logs from a specified Docker container.

Read-only operation with no side effects.

safe

list-containers

Lists all Docker containers and their status.

Read-only operation with no side effects.

Safety Assessment

This MCP server provides useful Docker management capabilities but requires careful consideration due to its ability to create, deploy, and delete containers. It's relatively safe when used for monitoring and log retrieval, but risky when deploying complex Compose stacks or creating containers without proper resource constraints.

Requires local Docker Engine, limiting external exposure.
Tool-based access control limits the scope of operations.
Clear tool definitions with specific parameters.
No direct internet access from the MCP server itself.

Can create, deploy, and delete containers, posing a risk of unintended consequences.
No built-in sandboxing; relies on Docker's container isolation.
Lack of resource limits can lead to resource exhaustion.
Compose deployments can introduce complex configurations with potential vulnerabilities.