red-team

Review artifacts adversarially with iterative attack, fix, and re-review loops.

10|2|Updated Feb 23, 2026
One-click install
npx skills add https://github.com/raddue/crucible --skill red-team-raddue
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: red-team
Source: https://github.com/raddue/crucible/tree/main/skills/red-team
Command: npx skills add https://github.com/raddue/crucible --skill red-team-raddue

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

This Skill provides a rigorous adversarial review process to identify critical flaws, hidden risks, and potential failure points in your designs, plans, code, or documentation before they cause problems.

Core Features & Use Cases

  • Automated Adversarial Testing: Dispatches a Devil's Advocate subagent to actively attack the artifact.
  • Iterative Improvement: The process loops, fixing issues and re-reviewing with fresh eyes until the artifact is clean or stagnation is detected.
  • Use Case: After completing a complex design document, use this Skill to have an AI act as a devil's advocate, finding potential integration issues, security vulnerabilities, or overlooked edge cases that human reviewers might miss.

Quick Start

Use the red-team skill to adversarially review the attached design document.

Frequently Asked Questions about red-team

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

FAQPage Schema
How do I perform an adversarial review of a design document?

Adversarial review of a design document is performed by dispatching a Devil's Advocate subagent to actively attack the artifact. The process iterates through fixing issues and re-reviewing until the document is clean or stagnation is detected.

What is automated red team testing for code and documentation?

Automated red team testing identifies critical flaws, hidden risks, and failure points in code or documentation. It uses a subagent to act as a devil's advocate, finding integration issues and security vulnerabilities human reviewers might miss.

Can I use adversarial testing to find security vulnerabilities in implementation plans?

You can use adversarial testing to find security vulnerabilities in implementation plans by having an AI attack the plan. The review loops through attack, fix, and re-review cycles to ensure robustness and identify critical flaws.

What is the best way to identify critical flaws in artifacts before deployment?

The best way to identify critical flaws in artifacts is rigorous adversarial review. A Devil's Advocate subagent attacks the artifact and iterates through fix and re-review loops, detecting potential failure points before they cause problems.

How does the review process handle stagnation during risk assessment?

The review process handles stagnation during risk assessment through built-in stagnation detection and escalation. This ensures the iterative attack, fix, and re-review loop terminates effectively when no further improvements are found.