adversarial-spec

Community

Refine specs via adversarial, multi-LLM debate.

Authorzscole
Version1.0.0
Installs0

System Documentation

What problem does it solve?

The adversarial-spec skill provides a structured workflow to converge on robust product specifications by simulating debates among multiple language models, surfacing gaps, edge cases, and biases that a single model would miss.

Core Features & Use Cases

  • Multi-model critique rounds for PRD and Technical Specifications to drive comprehensive requirements.
  • Deterministic outputs via standardized prompts, cost tracking, diffing, and session support.
  • Integrates with CLI tools and Bedrock/CLI providers to expose diverse perspectives and robust convergence.
  • Document evolution: stepwise revisions annotated with model feedback and Claude's critiques.
  • Supports final handoff to product, design, or engineering teams with a converged spec and traceable decisions.

Quick Start

  1. Prepare your initial concept or draft.
  2. Run a debate using the debate CLI, selecting available models, and specify document type (prd or tech).
  3. Review model feedback, apply changes, and iterate until all participants agree; export final spec.

Dependency Matrix

Required Modules

litellm

Components

scripts

💻 Claude Code Installation

Recommended: Let Claude install automatically. Simply copy and paste the text below to Claude Code.

Please help me install this Skill:
Name: adversarial-spec
Download link: https://github.com/zscole/adversarial-spec/archive/main.zip#adversarial-spec

Please download this .zip file, extract it, and install it in the .claude/skills/ directory.
View Source Repository

Agent Skills Search Helper

Install a tiny helper to your Agent, search and equip skill from 223,000+ vetted skills library on demand.