llm-council

Coordinate multiple LLMs through anonymous peer review and synthesis.

5|Updated Apr 15, 2026
One-click install
npx skills add https://github.com/47network/Sven --skill llm-council-47network
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llm-council
Source: https://github.com/47network/Sven/tree/main/skills/ai-agency/llm-council
Command: npx skills add https://github.com/47network/Sven --skill llm-council-47network

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Provides governance-aware, high-quality answers by coordinating multi-model deliberation and a structured synthesis process.

Core Features & Use Cases

  • Parallel deliberation across multiple LLMs to surface diverse perspectives.
  • Anonymous peer review and ranking to improve response quality and safety.
  • Configurable council composition with final synthesis by a chairman model, plus usage statistics and cost-awareness.

Quick Start

Submit a query to the council to receive a synthesized, peer-reviewed answer.

Frequently Asked Questions about llm-council

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

FAQPage Schema
How does multi-model deliberation improve LLM answer quality?▼

Multi-model deliberation improves answer robustness by running parallel queries across multiple LLMs, applying anonymous peer review to rank responses, and using a synthesis stage to finalize the output.

What is the best way to run anonymous peer review for LLM outputs?▼

The best way to run anonymous peer review for LLM outputs is to use a configurable council composition that evaluates responses anonymously, ranks them for safety and quality, and synthesizes the final answer.

Do I need a specific runtime LLM provider to use a model council?▼

Yes, you need a runtime LLM provider configured to supply the multiple models required for the council deliberation, peer review, and final synthesis process.

How do I configure council composition for multi-model synthesis?▼

You configure council composition by selecting the participating LLMs and specifying a chairman model responsible for the final synthesis, while the system validates your configuration and returns actionable errors if setup fails.

Can I track token usage and cost across multiple LLMs during deliberation?▼

Yes, you can track token usage and cost across multiple LLMs during deliberation, as the system provides usage statistics and cost-awareness alongside the final synthesized response.

What happens if the council configuration validation fails?▼

If the council configuration validation fails, the system stops the deliberation process and returns actionable errors to help you correct the setup before retrying the query.