# MCP SSH Manager — Remote Ops via Claude/Codex

> MCP SSH Manager is an MCP server that lets Claude Code and OpenAI Codex manage SSH sessions: run commands, sync files, and automate DevOps routines.

## Install

Merge the JSON below into your `.mcp.json`:

## Quick Use

```bash
# install
npm install -g mcp-ssh-manager

# add to Claude Code (example: from source path)
claude mcp add ssh-manager --scope user node /path/to/mcp-ssh-manager/src/index.js
```

## Intro

MCP SSH Manager is an MCP server that lets Claude Code and OpenAI Codex manage SSH sessions: run commands, sync files, and automate DevOps routines.

- **Best for:** DevOps automation, incident response, and remote runbooks (authorized only)
- **Works with:** Node.js + npm/npx; Claude Code; OpenAI Codex; SSH + rsync
- **Setup time:** 10–25 minutes

## Practical Notes

- GitHub: 199 stars · 35 forks; pushed 2026-05-07 (verified via GitHub API).
- README quick start includes `npm install -g mcp-ssh-manager` and `claude mcp add ... --scope user/project` options.
- README lists 37 tools and an optional auto-approve list for reducing repeated confirmations.

## Main

Treat SSH as a high-risk capability:

- Start with a **read-only** runbook (health checks, log tail, disk usage) and review the agent’s proposed commands.
- Use least-privilege SSH keys and separate “automation” accounts from human admin accounts.
- Keep an allowlist of hosts; avoid letting the agent accept arbitrary hostnames or copy secrets back into chat logs.

When it works well, the biggest win is consistency: every incident gets the same baseline checks before you escalate.

### FAQ

**Q: Can I use it without Claude Code?**
A: Yes—it's an MCP server; any MCP client that can spawn Node processes can use it (per README).

**Q: Should I enable auto-approve?**
A: Only for low-risk, read-only tools; keep write actions behind manual review.

**Q: How do I share configs across a team?**
A: README shows using `--scope project` to create a project-level `.mcp.json`.

## Source & Thanks

> Source: https://github.com/bvisible/mcp-ssh-manager
> License: MIT
> GitHub stars: 199 · forks: 35

---

<!-- ZH -->

## 快速使用

```bash
# 安装
npm install -g mcp-ssh-manager

# 添加到 Claude Code（示例：从源码路径运行）
claude mcp add ssh-manager --scope user node /path/to/mcp-ssh-manager/src/index.js
```

## 简介

MCP SSH Manager 是 MCP server，可让 Claude Code / OpenAI Codex 管理多台 SSH：执行命令、传文件/同步、做备份与健康检查等运维自动化，并提供 npm 一键安装。

- **适合谁：** 运维自动化 / 故障处理 / 远程 runbook（仅限授权）
- **可搭配：** Node.js + npm/npx；Claude Code；OpenAI Codex；SSH + rsync
- **准备时间：** 10–25 分钟

## 实战建议

- GitHub：199 stars · 35 forks；最近更新 2026-05-07（GitHub API 验证）。
- README quick start 包含 `npm install -g mcp-ssh-manager`，并提供 `--scope user/project` 的安装方式。
- README 提到 37 个工具，并提供可选 auto-approve 配置减少重复确认。

## 主要内容

把 SSH 当作高风险能力来设计：

- 先从 **只读 runbook** 起步（健康检查、日志查看、磁盘用量），并复核 agent 生成的命令。
- SSH key 最小权限，区分“自动化账号”和人类管理员账号。
- 主机做白名单，不要让 agent 接受任意 hostname，也不要把密钥/敏感配置回传到对话里。

真正的收益是“可重复”：每次故障先跑同一套基线检查，再决定是否升级处理。

### FAQ

**不使用 Claude Code 也能用吗？**
答：可以。它是 MCP server；任何能启动 Node 进程的 MCP 客户端都可用（见 README）。

**要不要开 auto-approve？**
答：只对低风险/只读工具开；写操作建议保留人工复核。

**团队怎么共享配置？**
答：README 展示 `--scope project` 生成项目级 `.mcp.json` 的方式。

## 来源与感谢

> Source: https://github.com/bvisible/mcp-ssh-manager
> License: MIT
> GitHub stars: 199 · forks: 35


---
Source: https://tokrepo.com/en/workflows/mcp-ssh-manager-remote-ops-via-claude-codex
Author: MCP Hub