Can I target specific parts of my project with Gemini?

Yes, you can use the --include-directories flag to specify exactly which folders or files Gemini should analyze, reducing noise and optimizing processing time.

How does the subagent delegation pattern work?

It uses Claude's Task tool to spawn a subagent that runs the Gemini query, asks for evidence, verifies the findings, and then provides a concise summary to the main agent.

Which Gemini models are supported by this skill?

The skill supports Gemini 2.5 (Flash, Pro, Flash-Lite) and Gemini 3 (Flash-Preview, Pro-Preview) models to balance speed, cost, and reasoning power.

When should I use Gemini instead of Claude?

Use Gemini for tasks requiring massive context windows (over 100k tokens), multi-file architecture reviews, or when you need a second opinion from a different model on critical analysis.

Do I need a Google Cloud account to use this?

Yes, you must have the Gemini CLI installed and configured with valid Google Cloud credentials or an API key to access the models.

Gemini CLI Integration

Name: Gemini CLI Integration
Author: fprochazka

byfprochazka

•

Developer Tools

Integrates Google’s Gemini models into Claude Code to handle massive context windows and deep architectural analysis.

About

This skill enables Claude Code to delegate complex tasks to Google’s Gemini CLI, leveraging context windows up to 1 million tokens for tasks that exceed Claude’s native limits. It is designed for analyzing large codebases, providing second opinions on complex logic, and performing multi-file architectural reviews. By utilizing a subagent delegation pattern, it ensures results are verified and summarized effectively, allowing developers to harness the power of Gemini 2.5 and 3 models directly within their Claude-assisted workflows for superior reasoning and broad codebase understanding.

Key Features

Multiple output formats including JSON, text, and streaming
Support for Gemini 2.5 and Gemini 3 Flash and Pro models
1 GitHub stars
Configurable directory inclusion for targeted project scanning
Automated subagent delegation for iterative verification
Massive 1M+ token context window for large codebase analysis

Use Cases

Getting a second opinion from Gemini on critical security or logic issues
Processing massive log files or documentation sets that exceed standard context limits
Analyzing entire repositories to map complex architecture and dependencies

About

Key Features

Multiple output formats including JSON, text, and streaming
Support for Gemini 2.5 and Gemini 3 Flash and Pro models
1 GitHub stars
Configurable directory inclusion for targeted project scanning
Automated subagent delegation for iterative verification
Massive 1M+ token context window for large codebase analysis

Use Cases

Getting a second opinion from Gemini on critical security or logic issues
Processing massive log files or documentation sets that exceed standard context limits
Analyzing entire repositories to map complex architecture and dependencies