M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
S

SSH Manager MCP Server

Community
SSH remote server management via Model Context Protocol
How to install ↓ Compare GitHub source repository ↗
Category
Dev Tools #227 of 438
Stars
★ 502 Very popular
Transport
stdio (local process)
Runtime
Node.js 18+
Credentials
No credential needed
License
MIT
Last commit
Tools
36
52FMRS · D

This server provides a broad feature set for comprehensive SSH management via MCP, suitable for teams needing deep DevOps automation. While security modes and audit logs are offered, users must still carefully manage credentials and permissions.

Strongest · Documentation 15/20 Weakest · Reliability 5/20

Reliability
5/20
Security and permissions
9/20
Maintenance
11/20
Documentation
15/20
Setup experience
12/20
Why each score
Reliability 5/20
Only the README and repository metadata were supplied; no source files, package.json, CI workflows, or committed tests are available. It is impossible to verify startup, the MCP init and tool-list handshake, or whether the 37 declared tools match real behavior. The README is untrusted evidence, and uncertainty lowers the score.
Security and permissions 9/20
No explicit red line is triggered: example passwords are placeholders, and the README claims shell-injection fixes, audit logging, and readonly/restricted modes. However, the default unrestricted mode, plaintext credentials in .env/TOML, a very broad remote-command surface, and the lack of source-level verification justify deductions.
Maintenance 11/20
The README shows an MIT license, version 3.7.0, a recent changelog, and security fixes, indicating apparent ongoing maintenance. But there is no verifiable commit history, issue-response trail, dependency-update evidence, or a clear security-response channel. Stars are a discovery signal only and do not add points.
Documentation 15/20
The README is extensive: install steps, authentication, tool parameters, security modes, troubleshooting, known limitations, and client configuration examples for Claude Code and Codex are all present. However, the referenced docs/ and test files cannot be verified from the supplied material, so full marks are not justified.
Setup experience 12/20
The README gives clear npm/npx install instructions, Claude Code and Codex client setup examples, and .env/TOML configuration samples. Nevertheless, the static calibration requires verifiable execution evidence, which is absent; setup also depends on Node, Bash, and optional rsync/sshpass, so the score is capped below 15.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessReads local filesWrites / deletes local filesRuns commands or codeUses the networkConnects to a database

Best for

  • Developers and DevOps engineers using Claude Code or OpenAI Codex
  • Teams managing multiple servers who want to automate routine tasks
  • Sysadmins needing quick command execution, file transfer, or monitoring

Not for

  • Non-technical users, as SSH and command-line basics are required
  • Highly security-sensitive environments if misconfigured, as SSH credentials exposure is a risk
  • Use cases requiring agent installation on each server, as this uses direct SSH

Required permissions

  • Access to local SSH config and credentials (passwords, keys, ssh-agent)
  • Ability to execute remote commands, transfer files, manage databases and backups
  • Ability to create SSH tunnels and persistent sessions
  • Ability to read and modify local .env and TOML configuration files

Risks and side effects

  • Credentials may be stored in plaintext in config files, posing security risks if files are compromised
  • Executing arbitrary commands via ssh_execute can cause unintended changes or damage
  • SSH agent forwarding can expose local keys to remote hosts if used with excessive permissions
  • Unstable network connections may lead to command timeouts or failures

Setup

Before you start

Runtime：Node.js 18+

Other optional settings (7)
SSH_SERVER_CI_ALLOW_PATTERNS optional Example ALLOW_PATTERNS regex allowlist for restricted mode; optional.
SSH_MANAGER_PROFILE optional Selects alias/hook profile (default/frappe/docker/nodejs); optional.
SSH_CONFIG_PATH optional Path to a TOML SSH config file for Codex setups; optional.
MCP_SSH_MAX_OUTPUT_LENGTH optional Output truncation length (default 10000 chars); performance tuning.
MCP_SSH_DEFAULT_TIMEOUT optional Command timeout in ms (default 120000); optional.
MCP_SSH_COMPACT_JSON optional Set true to use compact JSON to save tokens; optional.
SSH_AUTH_SOCK optional Local ssh-agent socket set by your system; when present, agent auth is used automatically.
  1. Install Node.js v18 or higher and globally install the npm package: npm install -g mcp-ssh-manager.
  2. Add the MCP server to Claude Code: claude mcp add ssh-manager node /path/to/mcp-ssh-manager/src/index.js.
  3. For OpenAI Codex, run ssh-manager codex setup and follow prompts to configure TOML server config.
  4. Configure servers using .env or TOML format (use ssh-manager server add wizard or manually edit config files).
  5. Install optional dependencies if using certain features: rsync (file sync), sshpass (rsync with password auth).

Check that it works

Run 'claude mcp list' to confirm ssh-manager is registered, then ask your AI client to call ssh_list_servers — it should return the servers you configured.

Troubleshooting

  1. If MCP tools are not available, check that MCP is installed and restart the client
  2. If connection fails, test with ssh-manager server test [server_name], check network, firewall, and SSH service
  3. If permission denied, verify username and credentials, check SSH key permissions (chmod 600)
  4. If large outputs cause client hanging, set MCP_SSH_MAX_OUTPUT_LENGTH or limit output with tail/grep
  5. For Windows hosts, ensure platform=windows is set in config to avoid command wrapping issues

Things to try

Once connected, you can ask your AI assistant things like:

  • List all my configured SSH servers
  • Execute 'df -h' on the production server
  • Upload config. to production:/etc/app/config.
  • Run a health check on production showing CPU, memory and disk status

Tools 36

ssh_list_servers read-only
Lists all configured SSH servers with their details.
ssh_execute writes
Execute commands on remote servers. Supports specifying working directory.
ssh_upload writes
Upload files to remote servers.
ssh_download read-only
Download files from remote servers.
ssh_sync writes
Bidirectional file sync with rsync between local and remote.
ssh_deploy writes
Deploy files with automatic permission and backup handling.
ssh_execute_sudo destructive
Execute commands with sudo privileges, securely handling password.
ssh_alias writes
Manage server aliases. Actions: add, remove, list.
Show 28 more tools
ssh_command_alias writes
Manage command aliases for frequently used commands. Actions: add, remove, list, suggest.
ssh_hooks writes
Manage automation hooks for SSH operations. Actions: list, enable, disable, status.
ssh_profile writes
Manage configuration profiles for different project types. Actions: list, switch, current.
ssh_backup_create writes
Create backup of database or files on remote server. Supports MySQL, PostgreSQL, MongoDB, and files.
ssh_backup_list read-only
List all available backups on remote server.
ssh_backup_restore writes
Restore from a previous backup.
ssh_backup_schedule writes
Schedule automatic backups using cron.
ssh_health_check read-only
Perform comprehensive health check on remote server (CPU, memory, disk, network, etc.).
ssh_service_status read-only
Check status of services (nginx, mysql, docker, etc.).
ssh_process_manager destructive
List, monitor, or kill processes on remote server.
ssh_alert_setup writes
Configure health monitoring alerts and thresholds. Actions: set, get, check.
ssh_db_dump read-only
Create database dump/backup on remote server (supports MySQL, PostgreSQL, MongoDB).
ssh_db_import destructive
Import SQL dump or restore database on remote server.
ssh_db_list read-only
List databases or tables on remote server.
ssh_db_query read-only
Execute read-only SQL queries on remote database (only SELECT allowed).
ssh_connection_status read-only
Check SSH connection status and force reconnect.
ssh_tunnel_create writes
Create SSH tunnel (local/remote port forwarding or SOCKS proxy).
ssh_tunnel_list read-only
List active SSH tunnels.
ssh_tunnel_close writes
Close SSH tunnel.
ssh_session_start writes
Start a persistent SSH session.
ssh_session_exec writes
Execute command in a persistent session.
ssh_session_end writes
End a persistent SSH session.
ssh_group_execute writes
Execute commands on multiple servers simultaneously.
ssh_monitor read-only
Real-time monitoring of CPU, memory, disk, and network.
ssh_tail read-only
Tail remote file.
ssh_disk_usage read-only
Check disk usage.
ssh_network_stats read-only
Get network statistics.
ssh_uptime read-only
Get uptime of remote server.

Use cases

Manage multiple servers remotely through an AI assistant
Automate DevOps tasks like deployments and backups
Monitor server health and performance
Perform database operations and secure management
Create and use SSH tunnels via AI

Supported clients

Claude Code
Codex

Listed from the project's documentation, not tested by this site.

Overview

An MCP server that enables Claude Code and OpenAI Codex to manage multiple SSH connections. Execute commands, transfer files, manage databases, create backups, monitor health, and automate DevOps tasks across your servers — directly from your AI assistant.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 62.9k · Tools 2 Compare with this →
Compare with other MCP servers

Source revision 161138373f4e Data synced 2026-10-12 Read the FMRS scoring method