Adversarial Reviewer

Generate structured red-team prompts to critique code before release.

27|8|Updated Jan 16, 2026
One-click install
npx skills add https://github.com/Sounder25/Google-Antigravity-Skills-Library --skill adversarial-reviewer
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: Adversarial Reviewer
Source: https://github.com/Sounder25/Google-Antigravity-Skills-Library/tree/main/19_adversarial_reviewer
Command: npx skills add https://github.com/Sounder25/Google-Antigravity-Skills-Library --skill adversarial-reviewer

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Adversarial Reviewer helps teams switch from building to critiquing by generating structured, red-team style prompts that expose weak assumptions and edge cases in code or plans before release.

Core Features & Use Cases

  • Structured critique prompts: Produces a ready-to-use adversarial review request for a target file and mode.
  • Mode-specific focus: Supports Security, Performance, and Logic perspectives to guide targeted reviews.
  • Pre-release QA workflow: Integrates into the development pipeline to surface issues before shipping critical components like auth or payments.

Quick Start

Run the PowerShell script to generate a review prompt: use the prepare_review.ps1 script with -FilePath set to the target artifact and -Mode to one of Security, Performance, or Logic.

Frequently Asked Questions about Adversarial Reviewer

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

FAQPage Schema
How do I generate a red-team code review prompt before shipping?

Run the prepare_review.ps1 PowerShell script with the -FilePath parameter set to your target artifact and -Mode set to Security, Performance, or Logic. This generates a structured markdown prompt for critiquing code and identifying weaknesses before release.

What is adversarial code review and when do I need it for pre-release QA?

Adversarial code review is a red-team critique technique that exposes weak assumptions and edge cases in code or plans before release. Use it during pre-release QA to target critical components like authentication, data handling, and design decisions.

Can I focus static analysis on specific areas like security or performance?

Yes, you can focus static analysis by selecting a review mode. The tool supports Security, Performance, and Logic perspectives, allowing you to generate mode-specific critique prompts that guide targeted code reviews for your target file.

What is the best way to integrate red-team critique into a development pipeline?

The best way to integrate red-team critique is running the prepare_review.ps1 script to generate markdown prompts, then feeding those prompts into a verification workflow. This validates the generated critique and surfaces issues before shipping critical components like payments or auth.

Do I need PowerShell to generate structured critique prompts?

Yes, you need PowerShell to run the prepare_review.ps1 script, which is the required mechanism for generating structured critique prompts. You must specify the target file path and review mode to produce the markdown output for analyzing your code artifacts.