Provides remote machine control capabilities, eliminating SSH overhead for token-efficient system operations.
A Model Context Protocol (MCP) server that provides comprehensive remote machine control capabilities for Claude Code, eliminating SSH overhead and providing token-efficient system operations.
# Clone or download the server files
git clone <repository> mcp-remote-machine
cd mcp-remote-machine
# Run installation script
./install_server.sh
Add the MCP server to Claude Code:
# For local machine control
claude mcp add remote-machine "python3 /path/to/remote_machine_server.py --transport stdio"
# For remote machine (over network)
claude mcp add remote-machine-net "python3 /path/to/remote_machine_server.py --transport sse --host 0.0.0.0 --port 8765"
In Claude Code, you'll have access to these tools:
authenticate_sudo
- Authenticate for privileged operationsexecute_command
- Run shell commandsread_file
/ write_file
- File operationslist_directory
- Browse directoriesprocess_management
- Control processesservice_control
- Manage systemd servicespackage_management
- Install/remove packagessystem_info
- Get system statusFirst authenticate for sudo operations:
> Use authenticate_sudo with your password
> Read the contents of /etc/nginx/nginx.conf
> Write a new configuration to /etc/myapp/config.json
> List all files in /var/log/ including hidden files
> Get current system information including CPU and memory usage
> Install the htop package
> Restart the nginx service
> Show all running processes
> Execute "df -h" to show disk usage
> Run "systemctl status docker" with sudo privileges
Aspect | SSH Approach | MCP Approach |
---|---|---|
Tokens per operation | ~150-300 | ~50-150 |
Connection overhead | New connection each command | Persistent connection |
Data format | Text parsing required | Structured JSON |
Authentication | Per-command prompts | Session-based cache |
Error handling | Text parsing | Structured responses |
Type safety | None | Full type annotations |
āāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāā
ā Claude Code āāāāāŗā MCP Server āāāāāŗā Target System ā
ā ā ā ā ā ā
ā - Tool calls ā ā - Authentication ā ā - File system ā
ā - Structured ā ā - Command exec ā ā - Processes ā
ā responses ā ā - System APIs ā ā - Services ā
āāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāāā āāāāāāāāāāāāāāāāāāā
STDIO Mode (recommended for local/SSH):
python3 remote_machine_server.py --transport stdio
SSE Mode (for network access):
python3 remote_machine_server.py --transport sse --host 0.0.0.0 --port 8765
Edit your Claude Code MCP configuration:
{
"mcpServers": {
"remote-machine": {
"command": "python3",
"args": ["/path/to/remote_machine_server.py", "--transport", "stdio"],
"env": {
"PYTHONPATH": "/path/to/server"
}
}
}
}
"Sudo authentication required"
authenticate_sudo
tool first with your password"Command not found" errors
/usr/bin/systemctl
)Connection issues
Run server with debug logging:
PYTHONPATH=/path/to/server python3 remote_machine_server.py --transport stdio
Installing a package via SSH:
Tokens: ~280
Command: ssh user@host "sudo apt install -y htop"
Response parsing: Complex text parsing required
Installing a package via MCP:
Tokens: ~120
Tool call: package_management(action="install", package_name="htop")
Response: Structured JSON with success/error status
Reading a file via SSH:
Tokens: ~200
Command: ssh user@host "cat /etc/hostname"
Response: Raw text output
Reading a file via MCP:
Tokens: ~80
Tool call: read_file(file_path="/etc/hostname")
Response: Structured with content, size, modified date
MIT License - see LICENSE file for details.
Single tool to control all 100+ API integrations, and UI components
Bootstrap Model Context Protocol (MCP) servers and clients in TypeScript with best practices, examples, and proper tooling setup.
Query and analyze your Opik logs, traces, prompts and all other telemtry data from your LLMs in natural language.
Reference / test server with prompts, resources, and tools
MCP server for text-to-graphql, integrates with Claude Desktop and Cursor.
AI-powered SVG animation generator that transforms static files into animated SVG components using the Allyson platform
Manage Buildkite pipelines and builds.
Enable AI agents to secure code with Semgrep.
A server for Zep, a long-term memory store for AI applications, requiring a ZEP_API_KEY for access.
Flag features, manage company data, and control feature access using Bucket.