gemini-review

Automate code review with Gemini CLI and Claude verification.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/dbaek-star/claude-gemini-skills --skill gemini-review-dbaek-star
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gemini-review
Source: https://github.com/dbaek-star/claude-gemini-skills/tree/main/skills/gemini-review
Command: npx skills add https://github.com/dbaek-star/claude-gemini-skills --skill gemini-review-dbaek-star

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill enables automated code review by orchestrating Gemini CLI reviews and Claude verification to improve accuracy, consistency, and auditability in collaborative coding workflows.

Core Features & Use Cases

  • Two-stage verification: Gemini provides review feedback and Claude confirms, disputes, or partially agrees to reduce errors.
  • Multi-file and directory support: Reviews can target a single file or an entire directory, with sequential processing and aggregated reports.
  • Audit-ready artifacts: Saves logs and decisions under .gemini/ for traceability, reuse, and auditing.

Quick Start

To start a review, provide the target code or directory to evaluate. The skill creates a working workspace under .gemini/review, runs Gemini first, then Claude, and outputs a final report for review.

Frequently Asked Questions about gemini-review

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate code review using both Gemini and Claude?

Automated code review using Gemini and Claude works by running Gemini CLI first to generate feedback, then having Claude confirm, dispute, or partially agree to ensure accuracy. This two-stage verification reduces errors and provides structured, auditable review artifacts.

Can I review an entire directory of code files for bugs and security issues?

Multi-file and directory reviews are supported for bug detection, performance, security, and style assessment. The skill processes targeted files sequentially, aggregates findings, and outputs a final report combining Gemini feedback with Claude verification.

What is the best way to generate auditable artifacts from an automated code review?

Auditable code review artifacts are generated by saving logs, structured verdicts, and decisions under a .gemini/ working directory. This ensures traceability and reuse by recording the stepwise workflow, model fallback actions, and final review reports.

Does the two-stage model verification workflow support model fallback?

Two-stage model verification supports model fallback within its stepwise workflow. If the primary review model encounters issues, the workflow executes fallback procedures to maintain consistent processing and generate the final structured report.

How do I start a code review workflow for my codebase?

To start a code review workflow, provide the target code file or directory to evaluate. The skill creates a workspace under .gemini/review, runs the two-stage Gemini and Claude verification, and outputs a final report for your assessment.