icon for mcp server

Terminal Command Execution

STDIO

A secure terminal execution server implementing MCP with command control and security features.

MCP Terminal Server

A secure terminal execution server implementing the Model Context Protocol (MCP). This server provides controlled command execution capabilities with security features and resource limits.

Features

  • Command Execution: Execute shell commands with output capture and error handling
  • Security Controls: Restrict allowed commands and prevent command injection
  • Resource Controls:
    • Command timeouts
    • Maximum output size limits
  • MCP Protocol Support:
    • Standard MCP message format
    • Capability advertisement
    • Streaming output support

Development

Local Setup

# Clone the repository git clone https://github.com/RinardNick/mcp-terminal.git cd mcp-terminal # Create and activate virtual environment using uv uv venv source .venv/bin/activate # or .venv\Scripts\activate on Windows # Install development dependencies uv pip install -e ".[dev]"

Publishing to PyPI

# Build the package uv pip install build python -m build # Upload to PyPI uv pip install twine python -m twine upload dist/*

Testing with MCP Inspector

The MCP Inspector tool can be used to test the server implementation:

# Install inspector npm install -g @modelcontextprotocol/inspector # Test server npx @modelcontextprotocol/inspector python3 src/mcp_terminal/server.py --allowed-commands "python,pip,git,ls,cd"

Running Tests

# Run all tests pytest tests/ # Run specific test file pytest tests/test_terminal.py # Run with coverage pytest --cov=mcp_terminal tests/

Using with Claude Desktop

Once the package is published to PyPI:

  1. Install UV (if not already installed):

    pip install uv
  2. Install the Package using UV:

    uv pip install mcp-terminal
  3. Configure Claude Desktop: Edit your Claude Desktop config file (typically at ~/Library/Application Support/Claude/claude_desktop_config.json on macOS):

    { "mcpServers": { "terminal": { "command": "uv", "args": [ "pip", "run", "mcp-terminal", "--allowed-commands", "python,pip,git,ls,cd", "--timeout-ms", "30000", "--max-output-size", "1048576" ] } } }

Protocol Implementation

The server implements the Model Context Protocol (MCP) with the following capabilities:

Capabilities Advertisement

{ "protocol": "1.0.0", "name": "terminal", "version": "1.1.0", "capabilities": { "execute": { "description": "Execute a terminal command", "parameters": { "command": { "type": "string", "description": "The command to execute" } }, "returns": { "type": "object", "properties": { "exitCode": { "type": "number" }, "stdout": { "type": "string" }, "stderr": { "type": "string" }, "startTime": { "type": "string" }, "endTime": { "type": "string" } } } } } }

Message Format

Request:

{ "type": "execute", "data": { "command": "echo 'hello world'" } }

Response:

{ "type": "result", "data": { "command": "echo 'hello world'", "exitCode": 0, "stdout": "hello world\n", "stderr": "", "startTime": "2024-01-20T12:34:56.789Z", "endTime": "2024-01-20T12:34:56.790Z" } }

Error:

{ "type": "error", "data": { "message": "command not allowed" } }

Security Considerations

  1. Command Validation:

    • Only allowed commands can be executed
    • Shell operators are blocked
    • Command injection attempts are prevented
  2. Resource Protection:

    • Command timeouts prevent hanging
    • Output size limits prevent memory exhaustion
    • Error handling for all failure cases
  3. Best Practices:

    • Always set allowed-commands in production
    • Use conservative timeout and size limits
    • Monitor command execution logs

Contributing

  1. Fork the repository
  2. Create your feature branch (git checkout -b feature/amazing-feature)
  3. Commit your changes (git commit -m 'Add some amazing feature')
  4. Push to the branch (git push origin feature/amazing-feature)
  5. Open a Pull Request

License

This project is licensed under the MIT License - see the LICENSE file for details.

Be the First to Experience MCP Now