How is the winner of the debate determined?

The skill uses a zero-sum 'tug-of-war' scoring system where a judge evaluates arguments; a positive total score indicates the proposition is leading, while a negative score favors the opposition.

Can I stop a debate and finish it later?

Yes, the skill is fully resumable. It tracks the current phase and exchange number in a local markdown file, allowing you to pick up exactly where you left off.

Does it require external scripts to run?

It utilizes a companion python package called 'debate_ops' included in the skill directory to process state changes and argument files.

What does the Debate Orchestrator skill do?

It manages a formal debate workflow in Claude Code, coordinating 'debater' and 'judge' subagents to argue a specific motion through structured rounds.

What is the difference between opening and rebuttal exchanges?

Opening exchanges (Exchange 0) involve both sides producing three independent arguments simultaneously. Subsequent rebuttal exchanges focus on one side attacking or defending specific points from previous rounds.

Debate Orchestrator

Name: Debate Orchestrator
Author: urav06

byurav06

•

生産性とワークフロー

Orchestrates structured, multi-agent formal debates with automated reasoning, rebuttals, and independent judging.

Debate Orchestrator is a specialized skill for the Claude Code CLI that automates complex dialectical exchanges between AI agents. It utilizes a deterministic state machine to coordinate proposition and opposition subagents through structured opening statements and iterative rebuttal rounds. By managing parallel execution and a dedicated judge agent, it provides a robust framework for exploring multiple perspectives on a motion, tracking scores through a zero-sum 'tug-of-war' system, and maintaining session resumability for long-form computational argumentation.

主な機能

01State-aware orchestration via the debate_ops Python package

02Structured phase management for openings and rebuttal exchanges

03Deterministic state tracking for session reliability and resumability

04Parallel multi-agent execution for simultaneous debater responses

05Automated zero-sum scoring to track winning/losing sides

064 GitHub stars

ユースケース

01Exploring complex policy decisions through adversarial reasoning

02Testing the robustness of arguments for research or writing

03Automated dialectical analysis of philosophical or technical motions

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add urav06/dialectic debate-orchestrator

For use in Claude.ai and ChatGPT

Download Skill