Does this skill modify my source code?

No, the skill is explicitly configured to provide feedback only. It uses Gemini to analyze diffs and provide commentary without making any edits to your files.

Why should I use Gemini to review Claude's work?

Cross-model verification provides a 'second opinion' from a different AI architecture, which helps catch edge cases, style inconsistencies, or bugs that a single model might overlook.

What specific vulnerabilities can the security review detect?

The security focus can be configured to check for XSS vulnerabilities, SQL/command injection, sensitive data exposure, and authentication issues.

Can Gemini see my entire repository?

Gemini can read files in your workspace directory that are not excluded by .gitignore, and it can be directed to read specific files to provide broader context for a review.

Gemini Diff Review

Name: Gemini Diff Review
Author: robbyt

byrobbyt

•

コラボレーションツール

Provides automated, multi-perspective code reviews for git changes using Google's Gemini AI.

概要

Gemini Diff Review is a specialized Claude Code skill designed to provide an independent quality assurance layer by leveraging Gemini to audit code changes made by Claude or the developer. By piping git diffs directly to the Gemini model, this skill offers a 'second opinion' on staged changes, specific commits, or uncommitted work. It is particularly effective for identifying subtle logic errors, security vulnerabilities, and performance bottlenecks that might be missed during the initial development phase, ensuring high code quality through cross-model verification.

主な機能

Customizable prompts for identifying specific issues like XSS, SQL injection, and memory leaks
Automated feedback on git staged changes and specific commit SHAs
Targeted review modes for security, performance, and style consistency
Deep context analysis by incorporating full file contents into the review process
Non-destructive workflow that provides feedback without altering source code
31 GitHub stars

ユースケース

Conducting focused security audits on sensitive authentication or data handling logic
Verifying AI-generated code changes for logic errors before a final commit
Performing performance reviews to identify N+1 queries or inefficient algorithms

概要

主な機能

Customizable prompts for identifying specific issues like XSS, SQL injection, and memory leaks
Automated feedback on git staged changes and specific commit SHAs
Targeted review modes for security, performance, and style consistency
Deep context analysis by incorporating full file contents into the review process
Non-destructive workflow that provides feedback without altering source code
31 GitHub stars

ユースケース

Conducting focused security audits on sensitive authentication or data handling logic
Verifying AI-generated code changes for logic errors before a final commit
Performing performance reviews to identify N+1 queries or inefficient algorithms