Gemini MCP

#!/bin/bash echo "===================================" echo "MCP 服务诊断脚本" echo "===================================" echo "" # 1. 检查虚拟环境 echo "1. 检查虚拟环境..." if [ -f "venv/bin/python3" ]; then echo "✅ 虚拟环境存在" echo " Python 版本: $(venv/bin/python3 --version)" else echo "❌ 虚拟环境不存在" exit 1 fi echo "" # 2. 检查依赖 echo "2. 检查依赖安装..." venv/bin/pip list | grep -E "fastmcp|openai|requests|python-dotenv" | while read line; do echo " ✅ $line" done echo "" # 3. 检查代码语法 echo "3. 检查代码语法..." venv/bin/python3 -c "from gemini_mcp import mcp" 2>&1 if [ $? -eq 0 ]; then echo "✅ MCP 模块可以导入" else echo "❌ MCP 模块导入失败" exit 1 fi echo "" # 4. 测试 MCP 服务器启动 echo "4. 测试 MCP 服务器启动（5秒超时）..." timeout 5 venv/bin/python3 -m gemini_mcp.mcp_server 2>&1 | head -20 echo "" # 5. 检查配置文件 echo "5. 检查配置文件..." if [ -f "$HOME/.cursor/mcp.json" ]; then echo "✅ Cursor 配置文件存在" echo " 路径: $HOME/.cursor/mcp.json" echo " 内容预览:" cat "$HOME/.cursor/mcp.json" | python3 -m json.tool | head -20 else echo "❌ Cursor 配置文件不存在" fi echo "" # 6. 检查工具注册 echo "6. 检查工具注册..." venv/bin/python3 -c " from gemini_mcp.mcp_server import send_images_to_gemini print(f'✅ 工具已注册: send_images_to_gemini') print(f' 类型: {type(send_images_to_gemini)}') " echo "" echo "===================================" echo "诊断完成" echo "===================================" echo "" echo "如果所有检查都通过但仍无法使用：" echo "1. 重启 Cursor" echo "2. 检查 Cursor 是否支持 MCP（需要最新版本）" echo "3. 查看 Cursor 的开发者控制台是否有错误信息"