Claude Council Orchestrator
AndreasAnadiotis/claude-council-orchestrator
Análise estratégica multi-perspectiva com síntese automática de consenso
Ver no GitHubFramework multi-agente avançado que executa avaliações paralelas de ideias através de três personas de IA customizadas (Otimista, Advogado do Diabo, Analista Neutro). Sintetiza perspectivas conflitantes em um dashboard executivo acionável, ideal para stress-test de modelos de negócio e estratégias operacionais.
Instalação
npx convez-ai AndreasAnadiotis/claude-council-orchestrator1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.
Métricas
Casos de uso
- Validação de modelos de negócio sob múltiplas ópticas
- Avaliação de risco e identificação de pontos cegos em estratégias
- Tomada de decisão executiva com análise de trade-offs estruturada
- Pre-mortem automático e planejamento de contingência
Veredito

Solução robusta de orquestração multi-agente que implementa um genuíno 'conselho' de perspectivas conflitantes para análise estratégica. Exige profundo entendimento de prompt engineering e arquitetura de agentes, mas entrega valor comprovado em decisões complexas e de alto risco.
README
A tradução para o português deste README ainda está na fila.
Multi-Agent Council Orchestrator for Claude Code
An advanced prompt-engineering and multi-agent framework built on top of the Claude Code CLI. This project configures a parallel evaluation protocol that stress-tests business models, operational strategies, and product ideas from three distinct, rigorous perspectives simultaneously.
🚀 How It Works
When triggered by the custom phrase council discuss [idea], the framework spins up three independent sub-agents in parallel to evaluate the proposal before synthesizing a final executive dashboard.
The Council Members:
- The Optimist Maximalist: Evaluates best-case trajectories, market ceilings, and high-value flywheels.
- The Devil's Advocate: Conducts pre-mortems, uncovers blind spots, and highlights structural vulnerabilities.
- The Neutral Analyst: Grounded anchor focused on data verification, trade-offs, and empirical uncertainties.
📂 Architecture
CLAUDE.md- The master orchestration layer and system rules.agents/optimist-maximalist.md- Persona configuration for uppercase potential.agents/devils-advocate.md- Persona configuration for risk mitigation.agents/neutral-analyst.md- Persona configuration for objective trade-off mapping.
📊 Sample Output: Council Consensus Dashboard
At the conclusion of every analysis, the orchestrator outputs a structured summary:
- Agent Recommendations: A 2-sentence distillation of each agent's independent stance.
- The Convergence Point: The core ground truth where all perspectives align.
- Actionable Next Steps: 4 immediate, sequential execution steps.
- Alternative Pivot: A lower-risk alternative path surfaced during the friction round.