do-red-team

Stress-tests ideas and plans using parallel adversarial agents producing steelman and counter-argument reports.

2|Updated May 25, 2026
One-click install
npx skills add https://github.com/edheltzel/Do-Skills --skill do-red-team-edheltzel
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: do-red-team
Source: https://github.com/edheltzel/Do-Skills/tree/main/skills/core/do-red-team
Command: npx skills add https://github.com/edheltzel/Do-Skills --skill do-red-team-edheltzel

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? People fall in love with their own plans and overlook flaws until those flaws surface in production, in the market, or in a critical meeting. This Skill attacks arguments, strategies, and plans deliberately and at volume so weak points surface while they are still cheap to fix. ## Core Features & Use Cases - Parallel Adversarial Analysis: Decomposes an argument into atomic claims and deploys up to 32 expert personas (engineers, architects, pentesters, interns) to attack each claim from distinct angles. - Steelman + Counter-Argument Output: Produces the strongest honest version of the argument and the strongest surviving counter-argument, each as 8 points of 12-16 words, ranked by severity. - Adversarial Validation (Battle of Bots): Generates competing proposals, applies brutal critique, and synthesizes a superior unified solution for design and architecture decisions. - Use Case: Ask it to red team a microservices migration plan, and it returns a steelman of the plan plus a devastating counter-argument exposing the hidden assumption that could collapse it. ## Quick Start Ask the agent to red team your plan or poke holes in your proposal, and it will run the ParallelAnalysis workflow to return a steelman and counter-argument.

Frequently Asked Questions about do-red-team

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

FAQPage Schema
How do I red team an idea or plan with AI?▼

Provide the argument, strategy, or plan and ask for a red team analysis. The skill decomposes it into atomic claims, deploys parallel adversarial personas to attack each claim, and returns a steelman plus the strongest counter-argument ranked by severity.

What is the difference between ParallelAnalysis and AdversarialValidation?▼

ParallelAnalysis stress-tests existing content by attacking it from many perspectives. AdversarialValidation produces new content through competition: multiple proposals are generated, brutally critiqued, then synthesized into a superior unified solution.

Can this skill find security vulnerabilities in my system?▼

No. This skill attacks ideas, arguments, and plans, not networks or code. It uses pentester-style personas for adversarial thinking about logic and assumptions, but it does not perform penetration testing or vulnerability scanning.

When should I not use adversarial validation?▼

Avoid it for simple decisions with obvious answers, time-critical situations, creative tasks with multiple valid outputs, or problems where expert consensus already exists. It trades speed for decision quality.

What output format does the red team analysis produce?▼

It produces a steelman of the argument and a counter-argument, each as 8 numbered points of 12-16 words. Findings are ranked by severity, target real weaknesses rather than strawmen, and include remediation paths.