SSH MCP Server
Enables remote command execution and file management over SSH and SFTP, supporting persistent connection storage and automated discovery for BinaryLane VPS servers.
README
SSH MCP Server
A Model Context Protocol (MCP) server for executing commands and managing files on remote servers via SSH.
Optionally auto-discovers BinaryLane VPS servers so they're available immediately.
Features
- Execute shell commands on remote servers
- Read files and list directories remotely
- Upload and download files via SFTP
- Persistent connections saved to config file across restarts
- BinaryLane auto-discovery of active servers
- Per-connection SSH key configuration
- Multiple connection sources with clear priority
Installation
npm install
npm run build
Configuration
Claude Desktop / Claude Code
Add to your Claude configuration:
{
"mcpServers": {
"ssh": {
"command": "node",
"args": ["/path/to/ssh-mcp/dist/index.js"],
"env": {
"BINARYLANE_API_TOKEN": "your-token-here"
}
}
}
}
The BINARYLANE_API_TOKEN is optional — only needed for auto-discovery of BinaryLane servers.
Config File
On first run, a config file is created at ~/.config/ssh-mcp/connections.json:
{
"defaultPrivateKeyPath": "~/.ssh/id_ed25519",
"binarylane": {
"enabled": true,
"apiToken": "your-token-here",
"defaultUsername": "root",
"defaultPrivateKeyPath": "~/.ssh/id_ed25519"
},
"connections": [
{
"name": "my-server",
"host": "10.0.0.5",
"port": 22,
"username": "deploy",
"privateKeyPath": "~/.ssh/deploy_key"
}
]
}
The BinaryLane API token can be set here or as an environment variable.
Environment Variable (Legacy)
Still supported for backward compatibility:
export SSH_CONNECTIONS='[
{"name": "web1", "host": "192.168.1.10", "port": 22, "username": "root", "privateKeyPath": "~/.ssh/id_rsa"}
]'
Connection Options
| Option | Required | Default | Description |
|---|---|---|---|
name |
Yes | - | Unique identifier for the connection |
host |
Yes | - | Server hostname or IP address |
port |
No | 22 | SSH port |
username |
Yes | - | SSH username |
privateKeyPath |
No | - | Path to SSH private key (supports ~ expansion) |
password |
No | - | SSH password (if not using key auth) |
If neither privateKeyPath nor password is provided, the server will try default SSH key locations (~/.ssh/id_ed25519, ~/.ssh/id_rsa).
Connection Priority
When multiple sources define the same connection name:
- Config file (highest) — manual connections from
~/.config/ssh-mcp/connections.json - Environment variable —
SSH_CONNECTIONS - BinaryLane auto-discovery (lowest) — active servers from the API
Available Tools
Connection Management
list_connections— List all available SSH connections with source infotest_connection— Test connectivity with latency reportingadd_connection— Add a connection (persisted to config file)remove_connection— Remove a manual connection from configrefresh_connections— Re-discover BinaryLane servers and reload all connectionsget_config— Show current configuration (tokens redacted)
Command Execution
run_command— Execute a shell command with configurable timeout
File Operations
read_remote_file— Read file contents via SFTPlist_remote_directory— List directory contents with metadataupload_file— Upload a local file to remote serverdownload_file— Download a remote file to local machine
Example Usage
# List all connections (shows source: config, env, or binarylane)
list_connections
# Test connection
test_connection connection="web1"
# Run a command
run_command connection="web1" command="uptime"
# Add a new persistent connection
add_connection name="staging" host="10.0.0.5" username="deploy" privateKeyPath="~/.ssh/deploy_key"
# Refresh after creating new BinaryLane servers
refresh_connections
Security Notes
- SSH keys are read from your local filesystem and never transmitted or stored by the MCP
- Passwords in config are only held in memory during operation
- The config file should be protected with appropriate filesystem permissions
- Commands are executed with the privileges of the SSH user
- Consider using dedicated SSH keys with limited permissions for automation
Development
npm run dev # Watch mode
npm run build # Compile TypeScript
npm start # Run compiled server
Project Structure
ssh-mcp/
├── src/
│ ├── index.ts # Server entry point
│ ├── ssh-client.ts # SSH connection manager
│ ├── config.ts # Persistent config file management
│ ├── binarylane.ts # BinaryLane auto-discovery
│ ├── tools.ts # MCP tool definitions
│ └── handlers.ts # Tool handler implementations
├── dist/ # Compiled JavaScript
├── package.json
└── tsconfig.json
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.