# LLM Council

> Run Fireworks-hosted open-weight model councils that compare responses and synthesize a final answer.

- Skill: `dair-ai/llm-council` (Agent Skill, multi-file: 3 files)
- Install (CLI): `npx skillmds@latest add dair-ai/llm-council`
- Raw SKILL.md: https://api.skillmd.com/api/skills/dair-ai/llm-council/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- License: MIT
- Author: dair-ai (https://skillmd.com/u/dair-ai)
- Updated: 2026-09-08
- Page: https://skillmd.com/skills/dair-ai/llm-council

---


# LLM Council (Fireworks AI)

## Detailed Guide

Read [the detailed guide](references/detailed-guide.md) before executing this skill. It retains the complete procedure and reference material. Treat its safety, prerequisites, and validation requirements as mandatory. For focused work, load the relevant sections; for end-to-end work, read the guide completely.

## When to Use

Use when this workflow matches the user request: Use this skill for its documented workflow.


_Source: [dair-ai/dair-academy-plugins](https://github.com/dair-ai/dair-academy-plugins) (MIT)._

This skill implements Karpathy's LLM Council concept where multiple open-weight LLMs deliberate on a query, powered entirely by Fireworks AI:

1. **Phase 1**: All models respond to the query independently (parallel)
2. **Phase 2**: Models rank each other's anonymized responses
3. **Phase 3**: A Chairman LLM synthesizes the final answer

All inference runs through **Fireworks AI** using open-weight models. The speed and pricing of Fireworks makes it practical to run multi-model deliberation that would be slow or expensive on other providers.

## CRITICAL RULES

1. **ALWAYS use AskUserQuestion** to let the user select council models (multiselect) and the Chairman model
2. **ALWAYS save raw responses to files** - never summarize or truncate API outputs
3. **ALWAYS show full transparency** - display all individual responses, all rankings, AND the final synthesis
4. **NEVER skip the ranking phase** - it is essential to the council deliberation process
5. **Read from files for display** - ensures content is shown unmodified
6. **ALWAYS display the final output to the user** after Phase 3 completes

## Limitations

- Requires the upstream tool, account, API key, or local setup when the workflow names one.
- Does not authorize destructive, production, paid, or external-message actions without explicit user approval.
- Validate generated artifacts or recommendations against the user's real sources before treating them as final.