debate-facilitator

Moderates structured multi-agent debates and produces minutes plus decision records.

Updated Dec 20, 2025
One-click install
npx skills add https://github.com/koala-man-64/helpful-scripts --skill debate-facilitator-koala-man-64
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: debate-facilitator
Source: https://github.com/koala-man-64/helpful-scripts/tree/main/exports/codex-skills/repo-local/debate-facilitator
Command: npx skills add https://github.com/koala-man-64/helpful-scripts --skill debate-facilitator-koala-man-64

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Multi-agent discussions often loop, lose track of objections, or end without a clear, traceable outcome. This Skill provides a neutral moderation process that turns agent disagreement into a documented decision, deferral, deadlock report, or escalation. ## Core Features & Use Cases - Structured Debate Lifecycle: Runs opening positions, clarifications, objections, rebuttals, synthesis, candidate proposals, and consensus checks across configurable rounds. - Consensus Tracking: Supports unanimity, consent, majority, supermajority, weighted, and chair-decides rules, with classified participant stances and blocking-objection handling. - Traceable Artifacts: Produces debate minutes, decision records, dissent documentation, deadlock reports, and optional JSON output with a decision quality score. - Use Case: When several agents disagree on a technical architecture choice, run a facilitated debate that captures each position, resolves blocking objections, and emits a formal decision record with dissent and action items. ## Quick Start Use the debate-facilitator skill to moderate a structured debate among the participating agents on this decision and produce minutes plus a final decision record.

Frequently Asked Questions about debate-facilitator

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

FAQPage Schema
How do I run a structured debate between multiple AI agents?▼

Provide the topic, decision required, participant list with roles, and a consensus rule, then start the debate. The facilitator runs opening positions, clarifications, objections, rebuttals, and consensus checks across up to five rounds by default.

What consensus rules does multi-agent debate moderation support?▼

Supported rules include unanimity, consent, majority, supermajority, weighted voting, and chair-decides. Consent is the default, meaning a proposal passes when no participant maintains a reasoned blocking objection.

What outputs does the debate facilitator produce?▼

It produces debate minutes, a decision record with rationale and alternatives, dissent documentation, action items, and a deadlock report when consensus fails. JSON output and an optional decision quality score are also available.

What happens when agents cannot reach consensus?▼

The facilitator runs focused resolution rounds on unresolved blockers up to the round limit. If deadlock remains, it emits a deadlock report listing blocking objections, areas of agreement, missing information, and a recommended next step such as escalation or experimentation.

When should I not use a structured multi-agent debate?▼

Avoid it for single-agent tasks, decisions with no genuine disagreement, or topics that are unsafe, underspecified, or out of scope. The facilitator terminates debates whose objective is invalid rather than forcing an artificial outcome.