Leverage Google Gemini's massive token window for analyzing large files and codebases through the Gemini CLI. Perfect for analyzing entire projects, processing large documents, or getting a second AI perspective.
Integrate Google Gemini's powerful analysis capabilities directly into your AI workflow through the Model Context Protocol (MCP). This skill enables you to leverage Gemini's massive token window for analyzing large files, entire codebases, and complex projects that might exceed other models' context limits.
This skill provides a bridge between Claude and Google Gemini through the Gemini CLI, allowing you to:
Before using this skill, ensure you have:
1. **Node.js** (v16.0.0 or higher) installed
2. **Google Gemini CLI** installed and configured ([installation guide](https://github.com/google-gemini/gemini-cli))
3. A valid Gemini API key configured in the Gemini CLI
```bash
claude mcp add gemini-cli -- npx -y gemini-mcp-tool
```
Verify installation by typing `/mcp` in Claude Code to confirm the gemini-cli MCP is active.
Add to your Claude Desktop config file:
**macOS**: `~/Library/Application Support/Claude/claude_desktop_config.json`
**Windows**: `%APPDATA%\Claude\claude_desktop_config.json`
**Linux**: `~/.config/claude/claude_desktop_config.json`
```json
{
"mcpServers": {
"gemini-cli": {
"command": "npx",
"args": ["-y", "gemini-mcp-tool"]
}
}
}
```
Then import to Claude Code:
```bash
claude mcp add-from-claude-desktop
```
Restart your terminal after configuration.
Use Gemini integration when:
Simply ask in natural language:
Use the `@` symbol to reference files and directories:
For safe code execution and testing, use sandbox mode:
Type `/gemini-cli` to see available slash commands:
```
"Use Gemini to analyze @. and provide a high-level architecture overview of this project. Focus on the main components, data flow, and any notable patterns."
```
```
"Ask Gemini to compare @src/oldImplementation.js with @src/newImplementation.js and explain the key differences and improvements."
```
```
"Use Gemini to analyze @logs/application.log and identify any errors, warnings, or unusual patterns. Summarize the findings."
```
```
"Use Gemini sandbox to create a Python script that reads @data.csv, performs basic data cleaning, and outputs summary statistics. Run it safely and show me the results."
```
```
"Ask Gemini to search for the latest security best practices for JWT authentication in Node.js applications and summarize the recommendations."
```
1. **Use File References**: Always use `@` syntax when referencing files to ensure Gemini can access the full content
2. **Be Specific**: Provide clear instructions about what you want Gemini to analyze or do
3. **Combine Perspectives**: Use Gemini for initial analysis of large codebases, then use Claude for detailed implementation
4. **Sandbox for Safety**: Always use sandbox mode when testing unknown code or scripts
5. **Model Selection**: Default model is `gemini-2.5-pro`, but you can specify different models if needed
If the MCP server doesn't appear:
1. Verify Node.js installation: `node --version`
2. Check Gemini CLI installation: `gemini --version`
3. Restart your terminal after configuration changes
4. Verify the MCP is active: type `/mcp` in Claude Code
For more help, visit the [full documentation](https://jamubc.github.io/gemini-mcp-tool/).
Leave a review
No reviews yet. Be the first to review this skill!
# Download SKILL.md from killerskills.ai/api/skills/gemini-mcp-integration/raw