LLM Council

Use for hard decisions, strategy calls, or "challenge my thinking / am I fooling myself" — spawns independent role-agents (each a distinct adversarial lens), collects their honest takes, then synthesizes as chairman. Anti-sycophancy by design. Triggers on "council", "bouscule-moi", "pressure-test", "decide X or Y", "pre-mortem", "red team my plan", or any high-stakes trade-off where the main model's agreement bias would hide the truth.

Install

skills CLI (Claude Code, Cursor, Codex and more)
npx skills add artguru-tech-team/claude-skills-public --skill llm-council
Manual install
git clone --depth 1 https://github.com/artguru-tech-team/claude-skills-public.git /tmp/claude-skills-public
mkdir -p ~/.claude/skills
cp -r /tmp/claude-skills-public/llm-council ~/.claude/skills/llm-council

Personal skills live in ~/.claude/skills. Use .claude/skills in a project folder to share a skill with your team.

From the SKILL.md

# LLM Council — multi-agent decision council

Adapté du *llm-council* de Karpathy (multi-modèle) → ici **multi-agent, single-model** dans Claude Code via le tool Agent. Plus besoin de plusieurs LLMs : plusieurs **agents indépendants**, chacun un rôle qui diverge structurellement, puis une synthèse de chairman.

## Pourquoi (le seul point qui compte)
Le modèle principal est **complaisant** : seul, il valide tes idées. Le council casse ça parce que les rôles sont des **subagents indépendants** qui (a) ne savent pas ce que tu/le chairman penses, (b) ont un mandat qui les pousse à diverger, (c) ont l'ordre explicite de ne PAS faire plaisir. La preuve que ça marche : le chairman se fait corriger par son propre council.

**Règle d'or : ne joue JAMAIS les rôles toi-même.** Si le modèle principal role-play les 5 voix, c'est toujours le même biais → théâtre. Utilise de vrais subagents (tool Agent).

## Process
1. **Brief factuel commun** — écris UNE fois les faits (chiffres réels, contraintes). Ground tous les agents dessus à l'identique. Si tu as des données (DB, métriques), tire-les AVANT — un council sur du narratif/vanity metrics produit du narratif.
2. **Fan-out** — spawn 4-5 agents EN PARALLÈLE (un seul message, plusieurs tool-calls Agent), `subagent_type: general-purpose`. Chacun : le brief + son rôle + l'ordre anti-complaisance + un format de sortie forcé. **Ne révèle pas ta reco ni le lean de l'utilisateur.**
Read the full skill on GitHub

Details

Skill name
llm-council
Path
llm-council/SKILL.md
Repo updated
11 November 2026

FAQ

What does the LLM Council skill do?

Use for hard decisions, strategy calls, or "challenge my thinking / am I fooling myself" — spawns independent role-agents (each a distinct adversarial lens), collects their honest takes, then synthesizes as chairman. Anti-sycophancy by design. Triggers on "council", "bouscule-moi", "pressure-test", "decide X or Y", "pre-mortem", "red team my plan", or any high-stakes trade-off where the main model's agreement bias would hide the truth.

How do I install the LLM Council skill?

skills CLI (Claude Code, Cursor, Codex and more): npx skills add artguru-tech-team/claude-skills-public --skill llm-council. Skills go in ~/.claude/skills (personal) or .claude/skills (project), and Claude loads them automatically when a task matches the description.

Is LLM Council popular?

It lives in artguru-tech-team/claude-skills-public, which has 10 GitHub stars and 0 forks.

Related skills

Amplitude Measurement Plan

10

artguru-tech-team/claude-skills-public

Amplitude measurement plan for a B2B two-sided marketplace.

Other

Appels Offres

10

artguru-tech-team/claude-skills-public

Use when responding to French public procurement tenders (appels d'offres, marchés publics, MAPA, DCE). Triggers on AO analysis, mémoire technique drafting, DCE review, CRF pricing, security questionnaire, candidature preparation. Also use when evaluating Go/No-Go on a tender.

Software Development

Brand Builder

10

artguru-tech-team/claude-skills-public

Use when building a complete brand identity for a new or existing brand — brand strategy, voice document, moodboard, logo concepts, color palette, and full AI-ready brand kit. Triggers on "build my brand", "create brand identity", "brand guidelines", "brand kit", or when a project needs visual + copy foundations from scratch.

Frontend & Design

Book to Skill

5

Londeren/claude-plugins

Use when the user wants to turn a knowledge source in Markdown (a book, manual, course, lecture notes, transcript) into a Claude skill, or to extract a method, methodology, or playbook from a large text into reusable agent instructions. Triggers on "make a skill from this book", "extract the methodology", "turn this into a playbook", "distill this source into a SKILL.md", "сделай скилл из книги", "перегони методичку в скилл", "извлеки методологию", "собери инструкцию по книге", "дистиллируй источник", and when the user brings a large .md with someone else's method and wants a working tool even

AI, LLMs & Agents

Dual Native Design

10

artguru-tech-team/claude-skills-public

Review and design software for both human users and AI agents — principles, checklists, and patterns for dual-native products.

AI, LLMs & Agents

Naiyue Thinking

5

tlzmw001/naiyue-skills

按耐阅的个人工程判断方式分析问题并组织行动。用于架构或方案取舍、项目学习与复杂机制解释、skill/prompt/agent 系统设计、代码审查、具有下游影响的决策,以及重复问题复盘。强调当前仓库事实、来源与产物边界、最小可用路径、AI 与确定性机制的职责分工,并保留用户对重要结论的最终确认。不用于纯事实查询、机械格式转换、无需判断的一步操作或闲聊。

AI, LLMs & Agents

Browse the most popular Claude skills, all ai, llms & agents skills, or write your own with the skills generator.