Enterprise DevOps MCP Server

中文 | English

Enterprise AI DevOps MCP Server for secure infrastructure automation with MCP, Docker, Kubernetes and SSH.

基于 MCP（Model Context Protocol） 的企业级 AI 运维 Agent Server：
让 AI Agent 在安全控制下管理 Linux、Docker、Kubernetes 与 SSH 基础设施。

AI Agent → MCP Server → Security Layer → Infrastructure

Architecture · Security · Contributing · Changelog · Release Notes

---

Why This Project

传统运维

痛点	表现
SSH 手动执行	工程师逐条敲命令，难规模化
权限不可控	模型或脚本一旦拿到 shell，几乎等同 root
缺少审计	谁改了什么、何时执行，事后难追溯

Engineer → SSH → Manual Command

AI DevOps（本项目）

AI Agent
   ↓
MCP Tool
   ↓
Security Layer
   ↓
Infrastructure

AI 不直接拥有服务器权限。 每次调用都经过 Permission Control、Execute Protection、Command Filtering 与 Audit Logging。

---

Features

Feature	Status
Linux Monitoring	✅
Docker Management	✅
Kubernetes	✅
SSH Automation	✅
Permission Control	✅
Execute Protection	✅
Audit Logging	✅
Docker Deployment	✅
GitHub Actions CI	✅

MCP Tools：17（System 7 · Docker 3 · Kubernetes 4 · SSH 3）

---

Security

企业级默认：只读运维。

EXECUTE_TOOLS_ENABLED=false

能力	说明
ReadOnly / Execute 分离	巡检与变更权限拆分
默认关闭执行权限	管理员显式开启后才可修改
Dangerous command filtering	建连前拦截危险 SSH 命令
Audit trail	Tool 调用可查询留痕

拦截示例

rm -rf /
   ↓
Blocked by security module
（不会发起远程 SSH）

详见：docs/security.md

---

Architecture

                 AI Agent
                    |
              MCP Protocol
                    |
        Enterprise DevOps MCP Server
                    |
 ------------------------------------------------
 |              |              |                |
System        Docker       Kubernetes        SSH
                    |
              Security Layer

详见：docs/architecture.md

---

Screenshots

1. MCP Server Connection
2. AI Infrastructure Health Check
3. Docker Container Inspection
4. MySQL Log Analysis
5. SSH Security Filter
6. Disk Cleanup Automation

更多：docs/screenshots/README.md

---

Quick Start

git clone https://github.com/zhifengjin050-arch/enterprise-devops-mcp-server.git
cd enterprise-devops-mcp-server

pip install -r requirements.txt
cp .env.example .env

python -m app.server

Cursor MCP

将 cwd 设为 YOUR_PROJECT_PATH（勿提交真实路径）：

{
  "mcpServers": {
    "enterprise-devops": {
      "command": "python",
      "args": ["scripts/run_devops_mcp.py"],
      "cwd": "YOUR_PROJECT_PATH",
      "env": {
        "FASTMCP_SHOW_SERVER_BANNER": "false",
        "EXECUTE_TOOLS_ENABLED": "false"
      }
    }
  }
}

示例：mcp_config_examples/cursor_mcp.json

Test

pytest

Docker

docker compose up -d --build

---

Project Structure

enterprise-devops-mcp-server/
├── README.md / README_EN.md
├── CONTRIBUTING.md / CHANGELOG.md
├── LICENSE
├── app/                 # MCP Server + Tools + Security
├── docs/                # Architecture / Security / Screenshots
├── examples/            # MCP client examples
├── tests/               # 225+ pytest cases
└── .github/workflows/   # CI

---

Contributing

见 CONTRIBUTING.md。

---

License

MIT

---

Disclaimer

开启执行权限后，本软件可控制服务器与容器。请先在非生产环境验证。