gsd-code-review

Analyze git diffs and generate consensus verdicts from a reviewer panel.

Updated May 11, 2025

Search Tags:#git #code review #AI #consensus #adversarial review

Ready to run

npx skills add https://github.com/rogsme/dotfiles --skill gsd-code-review-rogsme

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: gsd-code-review
Source: https://github.com/rogsme/dotfiles/tree/main/.config/gsd/skills/gsd-code-review
Command: npx skills add https://github.com/rogsme/dotfiles --skill gsd-code-review-rogsme

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill automates the code review process by using AI to perform adversarial code reviews, identifying potential bugs, security issues, and plan drift in the code.

Core Features & Use Cases

Adversarial Code Review: Multiple independent models review the code simultaneously.
Git Diff Analysis: Analyzes the actual git diff of a phase produced.
Review Panel: Sends the diff with context to a reviewer panel from a shared registry.
Consensus Verdict: Generates a consensus verdict based on reviewer findings.
Use Case: Before a code phase is executed, use this Skill to review the code and ensure it meets quality standards.

Quick Start

Run the command '/gsd-code-review N' to initiate a code review for phase N.

Frequently Asked Questions about gsd-code-review

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

FAQPage Schema

How do I automate code review using AI for git diffs?▼

Automate code review by analyzing git diffs and sending them to an AI reviewer panel for consensus. This process identifies bugs, security issues, and plan drift before executing a code phase.

What is adversarial code review and how does it work?▼

Adversarial code review uses multiple independent AI models to simultaneously review code. The models analyze git diffs with context and generate a consensus verdict based on their combined findings.

Do I need a shared registry to run AI-driven code reviews?▼

Yes, you need access to a shared registry of reviewers and their contracts. The Skill sends git diffs to this panel to obtain the consensus verdict required for the review.

How do I start a code review for a specific development phase?▼

Run the command '/gsd-code-review N' to initiate the code review for phase N. The Skill will analyze the git diff produced for that phase and send it to the reviewer panel.

Can AI code review detect security issues and plan drift?▼

Yes, the AI-driven adversarial review identifies potential bugs, security issues, and plan drift. It evaluates the git diff against the context to ensure the code meets quality standards.

What are the limitations of using AI consensus for git diff analysis?▼

The review relies on the shared registry of reviewers and their contracts. If the registry lacks diverse models or sufficient context, the consensus verdict may not catch all complex code issues.