ProcExecMCP
Stateless command execution and process management MCP server for Claude. Enables searching file contents, executing commands with safety limits, listing processes, and terminating processes.
README
ProcExecMCP
Stateless command execution and process management MCP server for Claude
ProcExecMCP enables Claude (acting in an architectural role) to search code content, execute commands, monitor processes, and terminate processes. The server exposes 4 MCP tools using FastMCP with Python 3.11+, psutil for cross-platform process management, and system ripgrep for file content search.
Features
- 🔍 search_file_contents: Search for patterns in file contents using ripgrep
- ⚡ execute_command: Execute commands safely with timeout and output limits
- 📊 list_processes: List running processes with filtering and sorting
- 🛑 kill_process: Terminate processes by PID
Security
- ✅ No shell injection (no
shell=True) - ✅ Mandatory timeouts on all operations
- ✅ Resource limits prevent memory exhaustion
- ✅ Path validation prevents traversal attacks
- ✅ Sanitized error messages
For comprehensive security documentation, see SECURITY_ARCHITECTURE.md.
Prerequisites
- Python 3.11+
- uv package manager (installation guide)
- ripgrep (rg) binary (installation guide)
- Claude Desktop or Claude for Windows
Quick Start
1. Install Dependencies
# Clone repository
git clone https://github.com/Positronikal/ProcExecMCP.git ProcExecMCP
cd ProcExecMCP
# Install with uv
uv sync
2. Verify ripgrep
# Check if ripgrep is installed
rg --version
If not installed:
- Windows:
winget install BurntSushi.ripgrep.MSVC - macOS:
brew install ripgrep - Linux:
sudo apt install ripgrep
3. Configure Claude Desktop
Edit your Claude Desktop configuration file:
- Windows:
%APPDATA%\Claude\claude_desktop_config.json - macOS:
~/Library/Application Support/Claude/claude_desktop_config.json - Linux:
~/.config/Claude/claude_desktop_config.json
Add ProcExecMCP to mcpServers:
{
"mcpServers": {
"procexec": {
"command": "uv",
"args": [
"--directory",
"/absolute/path/to/ProcExecMCP",
"run",
"procexec"
],
"env": {
"PROCEXEC_TIMEOUT": "30000",
"PROCEXEC_MAX_OUTPUT": "10485760",
"PROCEXEC_ENABLE_KILL": "true",
"PROCEXEC_RIPGREP_PATH": "/path/to/rg"
}
}
}
}
Windows Example:
{
"mcpServers": {
"procexec": {
"command": "C:\\Users\\YourName\\.local\\bin\\uv.exe",
"args": [
"--directory",
"D:\\dev\\ProcExecMCP",
"run",
"procexec"
],
"env": {
"PROCEXEC_RIPGREP_PATH": "C:\\Program Files\\ripgrep\\rg.exe"
}
}
}
}
4. Restart Claude Desktop
Close and reopen Claude Desktop to load the new MCP server.
5. Verify Installation
In Claude, try:
What MCP tools do you have available?
You should see all 4 ProcExecMCP tools listed.
Configuration
Environment Variables
- PROCEXEC_TIMEOUT: Command timeout in milliseconds (default: 30000, range: 1000-300000)
- PROCEXEC_MAX_OUTPUT: Maximum output size in bytes (default: 10485760 = 10MB)
- PROCEXEC_BLOCKED_PATHS: Comma-separated list of paths to block access
- PROCEXEC_ENABLE_KILL: Enable process termination tool (default: "true")
- PROCEXEC_RIPGREP_PATH: Full path to ripgrep binary (use if not in PATH)
Security Considerations
- Blocked Paths: Configure
PROCEXEC_BLOCKED_PATHSto protect sensitive directories - Process Termination: Disable with
PROCEXEC_ENABLE_KILL="false"if not needed - Timeouts: Set appropriate limits to prevent runaway commands
- Privileges: Avoid running Claude Desktop with elevated privileges unless necessary
Usage Examples
Search for Code Patterns
Search for all TODO comments in my Python project at /path/to/project
Execute Analysis Tools
Run 'pylint src/' in my project directory and show the results
Monitor Processes
List all Python processes sorted by memory usage
Terminate Stuck Processes
Terminate process 1234 gracefully
Development
For development setup, running tests, modifying the server, and building documentation, see USING.md.
Documentation
- Quick Start: See above
- Detailed Usage: USING.md
- Security Architecture: SECURITY_ARCHITECTURE.md
- API Documentation:
docs/directory (Doxygen generated) - Troubleshooting: BUGS.md
- Contributing: CONTRIBUTING.md
Troubleshooting
ripgrep not found
Ensure ripgrep is installed and in PATH, or set PROCEXEC_RIPGREP_PATH environment variable.
Timeout errors
Increase timeout via PROCEXEC_TIMEOUT environment variable (milliseconds).
Permission denied
Check file/directory permissions or configure PROCEXEC_BLOCKED_PATHS.
For more troubleshooting guidance, see BUGS.md.
License
This project is licensed under the GNU General Public License version 3 or any later version (GPLv3+). See COPYING.md for the full license text.
Contributing
Contributions are welcome! Please read CONTRIBUTING.md for guidelines.
Support
- Issues: GitHub Issues
- Security: See SECURITY.md for responsible disclosure
- Documentation: See
docs/for API reference
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。