Back to Details

gemini-deep-research

Compare original and translation side by side

🇺🇸

Original

English
🇨🇳

Translation

Chinese

Gemini Deep Research Skill

Gemini深度研究技能

Overview

概述

Claude can leverage Gemini Deep Research at gemini.google.com to conduct thorough, multi-source research on complex topics. Deep Research spends up to 20 minutes exploring the web, analyzing sources, and synthesizing findings into comprehensive reports.
Claude可以借助gemini.google.com上的Gemini深度研究功能，针对复杂主题开展全面的多来源研究。深度研究功能最多会花费20分钟浏览网页、分析来源，并将研究结果整合为详尽的报告。

Quick Install

快速安装

bash
curl -sSL https://canifi.com/skills/gemini-deep-research/install.sh | bash
Or manually:
bash
cp -r skills/gemini-deep-research ~/.canifi/skills/
bash
curl -sSL https://canifi.com/skills/gemini-deep-research/install.sh | bash
或者手动安装：
bash
cp -r skills/gemini-deep-research ~/.canifi/skills/

Setup

配置设置

Configure via canifi-env:
bash
undefined
通过canifi-env进行配置：
bash
undefined

First, ensure canifi-env is installed:

首先，确保已安装canifi-env：

curl -sSL https://canifi.com/install.sh | bash

curl -sSL https://canifi.com/install.sh | bash

canifi-env set GOOGLE_EMAIL "your-email@gmail.com"
undefined
canifi-env set GOOGLE_EMAIL "your-email@gmail.com"
undefined

Privacy & Authentication

隐私与认证

Your credentials, your choice. Canifi LifeOS respects your privacy.
您的凭证，由您掌控。 Canifi LifeOS尊重您的隐私。

Option 1: Manual Browser Login (Recommended)

选项1：手动浏览器登录（推荐）

If you prefer not to share credentials with Claude Code:
  1. Complete the Browser Automation Setup using CDP mode
  2. Login to the service manually in the Playwright-controlled Chrome window
  3. Claude will use your authenticated session without ever seeing your password
如果您不想与Claude Code共享凭证：
  1. 使用CDP模式完成浏览器自动化设置
  2. 在Playwright控制的Chrome窗口中手动登录服务
  3. Claude将使用您已认证的会话，全程不会获取您的密码

Option 2: Environment Variables

选项2：环境变量

If you're comfortable sharing credentials, you can store them locally:
bash
canifi-env set SERVICE_EMAIL "your-email"
canifi-env set SERVICE_PASSWORD "your-password"
Note: Credentials stored in canifi-env are only accessible locally on your machine and are never transmitted.
如果您愿意共享凭证，可以将其存储在本地：
bash
canifi-env set SERVICE_EMAIL "your-email"
canifi-env set SERVICE_PASSWORD "your-password"
注意：存储在canifi-env中的凭证仅能在您的本地设备访问，绝不会被传输至外部。

Capabilities

功能特性

  • Conduct in-depth research on any topic
  • Analyze multiple sources automatically
  • Generate structured research reports
  • Identify key themes and insights
  • Provide source citations
  • Compare different viewpoints
  • Summarize complex information
  • Answer multi-faceted questions
  • Research market trends and analysis
  • Investigate technical topics
  • 针对任意主题开展深度研究
  • 自动分析多来源信息
  • 生成结构化研究报告
  • 识别核心主题与关键见解
  • 提供来源引用
  • 对比不同观点
  • 总结复杂信息
  • 回答多维度问题
  • 研究市场趋势与分析
  • 调研技术主题

Usage Examples

使用示例

Example 1: Topic Research

示例1：主题研究

User: "canifi, research the best productivity systems for software engineers"
Claude: Opens Gemini, initiates Deep Research query.
        Waits for completion (up to 20 min), extracts findings.
        Stores synthesized report in Notion Research Hub.
用户："canifi，研究适合软件工程师的最佳生产力系统"
Claude：打开Gemini，发起深度研究查询。
        等待完成（最多20分钟），提取研究结果。
        将整合后的报告存储至Notion研究中心。

Example 2: Market Analysis

示例2：市场分析

User: "canifi, research the current state of AI in healthcare"
Claude: Launches Deep Research, monitors for completion.
        Returns comprehensive analysis with key players,
        trends, challenges, and opportunities.
用户："canifi，调研医疗领域AI的当前发展状况"
Claude：启动深度研究，监控完成进度。
        返回包含核心参与者、
        趋势、挑战与机遇的全面分析报告。

Example 3: Technical Investigation

示例3：技术调研

User: "canifi, research best practices for microservices architecture"
Claude: Initiates research, waits for completion.
        Provides detailed report on patterns, tools,
        case studies, and recommendations.
用户："canifi，研究微服务架构的最佳实践"
Claude：发起研究，等待完成。
        提供包含模式、工具、
        案例研究与建议的详细报告。

Example 4: Comparative Research

示例4：对比研究

User: "canifi, compare React vs Vue vs Svelte for my next project"
Claude: Runs Deep Research on all three frameworks.
        Delivers comparison table with pros, cons,
        use cases, and recommendations.
用户："canifi，对比React、Vue和Svelte，为我的下一个项目做选择"
Claude：对这三个框架运行深度研究。
        交付包含优缺点、
        适用场景与建议的对比表格。

Authentication Flow

认证流程

  1. Claude navigates to gemini.google.com via Playwright MCP
  2. Authenticates with GOOGLE_EMAIL if needed
  3. Handles 2FA if prompted (notifies user via iMessage)
  4. Maintains session for research operations
  1. Claude通过Playwright MCP导航至gemini.google.com
  2. 若需要，使用GOOGLE_EMAIL进行认证
  3. 若提示双因素认证，通过iMessage通知用户
  4. 维护会话以进行研究操作

Research Process

研究流程

1. Navigate to gemini.google.com
2. Click on Deep Research mode
3. Enter research query
4. Monitor progress (up to 20 minutes)
5. Extract and synthesize results
6. Store in appropriate Notion database
7. Report findings to user
1. 导航至gemini.google.com
2. 点击深度研究模式
3. 输入研究查询
4. 监控进度（最多20分钟）
5. 提取并整合结果
6. 存储至对应的Notion数据库
7. 向用户汇报研究结果

Selectors Reference

选择器参考

javascript
// Chat input
'[aria-label="Enter a prompt here"]' or 'rich-textarea'

// Deep Research button
'[aria-label="Deep Research"]' or button with research icon

// Submit button
'[aria-label="Send message"]'

// Response container
'.response-content' or '.model-response'

// Research progress indicator
'.research-progress'

// Sources panel
'.sources-panel'

// Copy response
'[aria-label="Copy"]'

// New chat
'[aria-label="New chat"]'
javascript
// 聊天输入框
'[aria-label="Enter a prompt here"]' or 'rich-textarea'

// 深度研究按钮
'[aria-label="Deep Research"]' or button with research icon

// 提交按钮
'[aria-label="Send message"]'

// 响应容器
'.response-content' or '.model-response'

// 研究进度指示器
'.research-progress'

// 来源面板
'.sources-panel'

// 复制响应
'[aria-label="Copy"]'

// 新聊天
'[aria-label="New chat"]'

Waiting Strategy

等待策略

javascript
// Deep Research can take up to 20 minutes
// Use periodic checking:
1. Launch research query
2. Wait 5 minutes (browser_wait_for)
3. Check for completion indicator
4. If not complete, wait another 5 minutes
5. Repeat until complete or 20 minute timeout
6. Extract results when done
javascript
// 深度研究最多需要20分钟
// 使用定期检查机制：
1. 发起研究查询
2. 等待5分钟（browser_wait_for）
3. 检查完成指示器
4. 若未完成，再等待5分钟
5. 重复直至完成或达到20分钟超时
6. 完成后提取结果

Error Handling

错误处理

  • Login Failed: Retry 3 times, notify user via iMessage
  • Session Expired: Re-authenticate automatically
  • Research Timeout: Notify user, offer to continue or restart
  • Query Rejected: Rephrase query, try again
  • Extraction Failed: Take screenshot, manually extract key points
  • Rate Limited: Wait and retry with backoff
  • 登录失败：重试3次，通过iMessage通知用户
  • 会话过期：自动重新认证
  • 研究超时：通知用户，提供继续或重启选项
  • 查询被拒绝：重新表述查询后重试
  • 提取失败：截取屏幕截图，手动提取关键点
  • 速率限制：等待后使用退避策略重试

Self-Improvement Instructions

自我优化说明

When you learn a better way to accomplish a task with Gemini Deep Research:
  1. Document the improvement in your response
  2. Suggest updating this skill file with the new approach
  3. Include specific query formulations that work better
  4. Note any UI changes affecting the workflow
当您发现使用Gemini深度研究完成任务的更优方法时：
  1. 在响应中记录该优化点
  2. 建议使用新方法更新此技能文件
  3. 包含效果更好的具体查询表述方式
  4. 记录影响工作流程的任何UI变更

Notes

注意事项

  • Deep Research is the PRIMARY research method for LifeOS
  • Always synthesize results before storing in Notion
  • Research can take 5-20 minutes; plan accordingly
  • Results include citations and source links
  • Research quality improves with specific, well-formed queries
  • Consider running in background agent for long research tasks
  • All research outputs should be stored in Notion Research Hub
  • Never use WebSearch for LifeOS research; always use Deep Research
  • 深度研究是LifeOS的主要研究方法
  • 存储至Notion前务必整合研究结果
  • 研究耗时5-20分钟，请合理规划时间
  • 结果包含引用与来源链接
  • 具体、表述清晰的查询能提升研究质量
  • 对于长时间研究任务，可考虑在后台Agent中运行
  • 所有研究输出均应存储至Notion研究中心
  • LifeOS研究绝不要使用WebSearch，务必使用深度研究