NM

neptaco/mcproc

开发工具

本页由 AI 翻译,技术细节请以原文为准。 查看原文

10 stars 5 forks 质量 72 趋势 22MIT

A Model Context Protocol (MCP) server for comfortable background process management on AI agents.

概览

A Model Context Protocol (MCP) server for comfortable background process management on AI agents. mcproc bridges the gap between AI agent development and traditional command-line workflows. It enables AI agents to manage long-running development processes (like dev servers, build watchers, etc.) while providing developers with full CLI access to monitor and control these same processes. Simple AI agent-launched processes are stateless and can't manage long-running processes effectively. mcproc solves this by: - : No more confusion about which agent or terminal is running what - all processes are centrally managed - : Logs are captured and stored, allowing AI agents to debug issues while reviewing logs from earlier - : Full CLI access means you're never locked out of your own development environment - 🔄 : Start and manage background processes from AI agents via MCP, then monitor them from your terminal - 👁️ : Processes started by AI agents are fully accessible via CLI and other...

README

以下为原文 README 节选,完整内容请查看原文。

mcproc

A Model Context Protocol (MCP) server for comfortable background process management on AI agents.

English | 日本語

Overview

mcproc bridges the gap between AI agent development and traditional command-line workflows. It enables AI agents to manage long-running development processes (like dev servers, build watchers, etc.) while providing developers with full CLI access to monitor and control these same processes.

Why mcproc?

Simple AI agent-launched processes are stateless and can’t manage long-running processes effectively. mcproc solves this by:

  • Unified Control: No more confusion about which agent or terminal is running what - all processes are centrally managed
  • Context Preservation: Logs are captured and stored, allowing AI agents to debug issues while reviewing logs from earlier
  • Developer-Friendly: Full CLI access means you’re never locked out of your own development environment

Key Features

  • 🔄 Unified Process Management: Start and manage background processes from AI agents via MCP, then monitor them from your terminal
  • 👁️ Cross-Environment Visibility: Processes started by AI agents are fully accessible via CLI and other agents, and vice versa
  • 📝 Intelligent Log Management: Capture, persist, and search process logs with powerful regex patterns
  • 📁 Project-Aware: Automatically groups processes by project context
  • 📊 Real-time Monitoring: Follow logs in real-time from CLI while AI agents manage the processes
  • 🛡️ XDG Compliant: Follows XDG Base Directory specification for proper file organization
  • ⚡ Wait-for-Log: Start processes and wait for specific log patterns to ensure readiness
  • 🔍 Advanced Search: Time-based filtering, context lines, and regex support for log analysis
  • 🧰 Toolchain Support: Execute commands through version managers (mise, asdf, nvm, rbenv, etc.)
  • 🧹 Clean Command: Stop all processes in a project with a single command
  • 🌲 Process Groups: Automatic cleanup of child processes when stopping parent processes

Installation

Using Homebrew (macOS and Linux)

# Add the tap
brew tap neptaco/tap

# Install mcproc
brew install mcproc

Build from source

Prerequisites

  • Rust toolchain (rustc, cargo)
  • protobuf compiler:
    • macOS: brew install protobuf
    • Linux: apt-get install protobuf-compiler
git clone https://github.com/neptaco/mcproc.git
cd mcproc
cargo build --release

# Install to PATH (optional)
cargo install --path mcproc

Usage

Setup as MCP Server

After installing mcproc, you need to register it as an MCP server with your AI assistant.

For Claude Code

# Register mcproc as an MCP server
在 GitHub 查看完整 README

安装

This server does not publish a one-line install command.

Open the repository installation guide 新手?先看 mcpskill.org 教程 →

配置

{ "mcpServers": { "mcproc": { "command": "mcproc", "args": ["mcp", "serve"] } } }