agent-council

Orchestrate multi-agent structured debates to analyze complex decisions and surface hidden assumptions.

40|6|Updated Jul 11, 2026

Search Tags:#architecture #multi-agent #debate #risk-assessment #decision-making #pydantic-ai

Ready to run

npx skills add https://github.com/magnus919/agent-skills --skill agent-council-magnus919

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: agent-council
Source: https://github.com/magnus919/agent-skills/tree/main/agent-council
Command: npx skills add https://github.com/magnus919/agent-skills --skill agent-council-magnus919

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pydantic-ai, pyyaml, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill resolves high-stakes, ambiguous questions by simulating a structured, multi-perspective debate, preventing cognitive biases like groupthink and positional commitment.

Core Features & Use Cases

Structured Debate Protocol: Executes a rigorous process including premortems, iterative cross-examination, and convergence-aware synthesis.
Convergence Diagnostics: Provides quantitative confidence dispersion and argument novelty metrics to determine if a decision is truly resolved.
Use Case: Use this to evaluate architectural trade-offs, such as choosing between database technologies, by spawning expert agents to stress-test assumptions and surface hidden risks before you commit.

Quick Start

Run the agent-council skill to debate the question of whether to migrate from SQLite to Postgres for the current service.

Frequently Asked Questions about agent-council

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

FAQPage Schema

How do I use multi-agent debate for architectural decision-making and risk assessment?▼

Multi-agent debate for architectural decision-making uses pydantic-ai to orchestrate adversarial agents that stress-test assumptions, run premortems, and cross-examine trade-offs to surface hidden risks before you commit.

What is convergence diagnostics in structured multi-agent debate?▼

Convergence diagnostics in structured multi-agent debate provides quantitative confidence dispersion and argument novelty metrics to determine whether a decision is truly resolved or requires further cross-examination.

Can I evaluate database migration trade-offs using pydantic-ai multi-agent orchestration?▼

Yes, you can evaluate database migration trade-offs by spawning expert agents with pydantic-ai to simulate a structured debate, stress-test assumptions, and surface hidden risks before committing.

Do I need Python 3.10 and pydantic-ai to run structured debate graphs?▼

Yes, you need Python 3.10 or higher and pydantic-ai installed to execute the structured debate graph and generate typed synthesis reports with convergence diagnostics.

How does structured debate prevent groupthink in complex architectural decisions?▼

Structured debate prevents groupthink in complex architectural decisions by simulating adversarial collaboration across multiple expert agents, forcing iterative cross-examination and premortems instead of positional commitment.

When should I not use multi-agent debate for decision-making?▼

You should avoid multi-agent debate for simple, low-stakes decisions where the overhead of spawning expert agents and running convergence diagnostics outweighs the benefit of surfacing hidden assumptions.