gsd-review

Automate multi-model adversarial review of software project phases.

Updated May 11, 2025
One-click install
npx skills add https://github.com/rogsme/dotfiles --skill gsd-review-rogsme
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gsd-review
Source: https://github.com/rogsme/dotfiles/tree/main/.config/gsd/skills/gsd-review
Command: npx skills add https://github.com/rogsme/dotfiles --skill gsd-review-rogsme

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill addresses the need for comprehensive plan reviews by leveraging the power of multiple AI reviewers to provide structured, adversarial feedback on software project phases.

Core Features & Use Cases

  • Cross-AI Review: Utilizes a panel of AI reviewers to provide structured feedback on implementation plans.
  • Adversarial Feedback: Each reviewer provides independent, adversarial feedback to catch blind spots and potential issues.
  • Integration: Automatically generates a review summary report that can be used to refine plans. Suitable for software development phases.

Quick Start

To review phase N of your project, simply run: gsd-review N

Frequently Asked Questions about gsd-review

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

FAQPage Schema
How do I get cross-AI adversarial feedback on software project phases?

Cross-AI adversarial review works by sending all relevant project documents to a predefined panel of external AI reviewers, which then independently analyze the plans and output a detailed consensus feedback report.

What's the best way to automate multi-model plan reviews for software development?

The best way to automate multi-model plan reviews is to trigger a command that distributes phase documents to a panel of AI reviewers, automatically generating a summary report to refine your implementation plans.

Do I need external AI reviewers with specific CLI support to run automated phase reviews?

Yes, you need external reviewers with specific CLI support to execute the automated phase review, as the process relies on these external models to independently generate and return adversarial feedback.

How does cross-AI adversarial feedback catch blind spots in implementation plans?

Cross-AI adversarial feedback catches blind spots by having multiple independent AI reviewers analyze the same software project phase documents simultaneously, identifying potential issues that a single reviewer might miss.

Can I use this automated review for specific phases of my software project?

Yes, you can review specific phases of your software project by passing the phase number to the command, which automatically gathers all relevant documents for that phase and sends them to the AI panel.