
Claude Code 到 Codex 合作的起点
概览
将AI助手整合进开发工作流程的全面指南。这些模板有助于建立不同AI模型与人类开发者之间一致且有效的协作。 - - Claude Code 及类似 AI 助手指南 - 根因分析方法论 - 常见错误及避免方法 - 多模型协作模式 - MCP 服务器集成(如有)- 项目特定配置模板 - - 多智能体协作框架 - 代理角色与优势 - 助理之间的交接协议 - 任务委托模式 - 协作工作流程 - - MCP 服务器设置指南 - 一行代码设置 - 添加其他 MCP 服务器 - 安全选项 - 故障排除 - - 简单MCP配置模板- - 环境变量模板 1. 在 Claude 中,询问:“你能使用 Codex 创建测试文件吗?”2. Claude 应该委派给 Codex 执行:- 创建test-codex-integration.txt,并写“来自 Codex 的问候!”
概述(AI 翻译)
将AI助手整合进开发工作流程的全面指南。这些模板有助于建立不同AI模型与人类开发者之间一致且有效的协作。
📚 文档结构
核心指南
-
CLAUDE.md - Claude Code 及类似 AI 助手指南
-
根本原因分析方法论
-
常见错误及避免方法
-
多模式协作模式
-
MCP服务器集成(如有)
-
项目专用配置模板
-
AGENTS.md - 多智能体协作框架
-
代理角色与优势
-
助理之间的交接协议
-
任务委派模式
-
协作工作流程
-
MCP_SETUP.md - MCP 服务器设置指南
-
一行法典设置
-
添加其他MCP服务器
-
安全选项
-
故障排除
配置示例
- mcp-config-example.json - 简单的MCP配置模板
- .env.example - 环境变量模板
🚀 快速启动
1.启用MCP集成(一个命令!)
2.快速烟雾测试(60秒)
立即测试积分:
- 在Claude中,问:“你能用Codex创建测试文件吗?”
本页由 AI 翻译,技术细节请以原文为准。
README
以下为原文 README 节选,完整内容请查看原文。
AI Assistant Documentation Hub
A comprehensive guide for integrating AI assistants into your development workflow. These templates help establish consistent, effective collaboration between different AI models and human developers.
📚 Documentation Structure
Core Guidelines
-
CLAUDE.md - Guidelines for Claude Code and similar AI assistants
- Root cause analysis methodology
- Common mistakes and how to avoid them
- Multi-model collaboration patterns
- MCP server integration (when available)
- Project-specific configuration template
-
AGENTS.md - Multi-agent collaboration framework
- Agent roles and strengths
- Hand-off protocols between assistants
- Task delegation patterns
- Collaboration workflows
-
MCP_SETUP.md - MCP server setup guide
- One-line Codex setup
- Adding other MCP servers
- Security options
- Troubleshooting
Configuration Examples
- mcp-config-example.json - Simple MCP config template
- .env.example - Environment variables template
🚀 Quick Start
1. Enable MCP Integration (One Command!)
# Install Codex CLI (choose your preferred method)
npm install -g @openai/codex # via npm
brew install --cask codex # via Homebrew (macOS)
# Sign in with your ChatGPT account (Plus/Pro/Team/Edu/Enterprise plans)
codex # then select "Sign in with ChatGPT"
# OR set your OpenAI API key for API key auth:
# export OPENAI_API_KEY=sk-...
# Add Codex MCP server to Claude (macOS/Linux)
claude mcp add codex -s user -e OPENAI_API_KEY=$OPENAI_API_KEY -- codex mcp -c model=gpt-5.3 -c model_reasoning_effort="high"
# Windows PowerShell variant
claude mcp add codex -s user -e OPENAI_API_KEY=$env:OPENAI_API_KEY -- codex mcp -c model=gpt-5.3 -c model_reasoning_effort="high"
# Restart Claude Desktop - Done! 🎉
2. Quick Smoke Test (60 seconds)
Test the integration immediately:
- In Claude, ask: “Can you work with Codex to create a test file?”
- Claude should delegate to Codex to:
- Create
test-codex-integration.txtwith “Hello from Codex!” - Show the file contents
- Delete the file
- Create
- Success = file created, shown, and removed with no other changes
3. Copy Templates to Your Project
# Clone or download these templates
cp -r docs/ai-assistants /path/to/your/project/docs/
# Customize for your project
cd /path/to/your/project/docs/ai-assistants
4. Configure for Your Project
Edit the configuration section in CLAUDE.md:
# In CLAUDE.md - Add your project details
project:
name: "Your Project Name"
type: "web|cli|library|service"
language: "TypeScript|Python|Go|etc"
conventions:
style_guide: "link-to-style-guide"
test_framework: "jest|pytest|etc"
min_coverage: 80
🤝 Collaboration Protocol
Working with AI Assistants
- Primary Orchestrator (Claude Code or similar)
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 →