RedTeam

Coordinate 32 agents to stress-test decisions with adversarial analysis.

28|5|Updated Feb 8, 2026
One-click install
npx skills add https://github.com/jgmontoya/shaka --skill redteam-jgmontoya
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: RedTeam
Source: https://github.com/jgmontoya/shaka/tree/main/defaults/system/skills/RedTeam
Command: npx skills add https://github.com/jgmontoya/shaka --skill redteam-jgmontoya

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provides structured, multi-perspective critique by coordinating 32 agents to aggressively challenge assumptions, surface blind spots, and produce high-quality counterpoints.

Core Features & Use Cases

  • Parallel critique: 32 agents (engineers, architects, pentesters, interns) examine strengths and weaknesses in parallel.
  • Structured outputs: Delivers a steelman and a counter-argument with concise reasoning.
  • FirstPrinciples integration: Utilizes FirstPrinciples/Deconstruct and FirstPrinciples/Challenge to surface hard constraints and hidden assumptions.

Quick Start

Summon the 32-agent parallel analysis to stress-test a given argument.

Frequently Asked Questions about RedTeam

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

FAQPage Schema
What is adversarial analysis for stress-testing decisions?

Stress-testing decisions involves coordinating 32 parallel agents, including engineers, architects, and pentesters, to aggressively challenge assumptions and surface blind spots. This adversarial analysis produces structured steelman arguments and robust counterpoints for architecture, product, and policy prompts.

How do I run multi-perspective critique on a product proposal?

To run multi-perspective critique on a product proposal, you summon the 32-agent parallel analysis. This process examines strengths and weaknesses in parallel, leveraging a 24-claim decomposition workflow to surface hard constraints and deliver concise counter-arguments.

Do I need the FirstPrinciples skill to perform parallel critique?

Yes, you need the FirstPrinciples skill to perform parallel critique. The RedTeam skill requires FirstPrinciples specifically for its Deconstruct and Challenge phases to surface hard constraints and hidden assumptions during the 24-claim decomposition workflow.

What is the best way to surface hidden assumptions in an architecture prompt?

The best way to surface hidden assumptions in an architecture prompt is using a 24-claim decomposition workflow. By applying FirstPrinciples Deconstruct and Challenge phases, 32 parallel agents can aggressively expose blind spots and generate structured counterpoints.

Can I use this adversarial analysis for policy prompts?

Yes, you can use this adversarial analysis for policy prompts. The 32-agent parallel critique is designed to apply to architecture, product, and policy prompts, effectively surfacing weaknesses, insights, and robust counterpoints across all these domains.

When should I not use a 32-agent parallel critique?

You should not use 32-agent parallel critique when lacking the required FirstPrinciples skill dependency, as it is mandatory for the Deconstruct and Challenge phases. The advanced implementation depth also makes it unsuitable for simple queries not requiring deep adversarial analysis.