What is the typical latency for a council response?

Standard team mode queries typically take 30-120 seconds, while the comprehensive debate mode can take 5-20 minutes depending on the complexity of the task and model response times.

What is the difference between Team Mode and Debate Mode?

Team Mode is faster and uses a single-pass 4-phase internal deliberation, while Debate Mode is a 4-round adversarial process that includes cross-critiques and a final audit verdict for maximum reliability.

How does the Council skill improve AI response quality?

By querying three independent flagship models and forcing them through a 4-phase deliberation or debate, the skill identifies hallucinations and highlights model-specific biases to reach a higher-quality consensus.

Does this skill work with my local files?

Yes, you can inject local file contents into the council's prompts by using the CONTEXT_FILES environment variable to provide domain-specific information.

Multi-Model Council Synthesis

Name: Multi-Model Council Synthesis
Author: twwoo0210

bytwwoo0210

0•

Data Science & ML

Synthesizes unified answers by orchestrating independent deliberations and adversarial debates between Claude, GPT-5 Codex, and Gemini Pro.

The Council skill empowers Claude Code to act as a primary orchestrator, querying three flagship AI models—Claude Opus 4.6, GPT-5.3 (Codex), and Gemini 3.1 Pro—to solve complex problems. It uses structured multi-phase prompting to simulate a professional team environment (Researcher, Analyst, Critic, Team Lead) or a high-intensity 4-round adversarial debate. This multi-perspective approach is designed to minimize model bias, identify edge cases, and provide a high-confidence consensus for critical technical decisions and architectural planning.

Key Features

01Simultaneous multi-model querying of Claude, GPT-5.3 Codex, and Gemini 3.1 Pro

02Automated synthesis highlighting consensus, divergence, and audited recommendations

034-Phase 'Team Deliberation' mode for structured internal reasoning

044-Round adversarial 'Deep Dive Debate' mode for high-stakes problem solving

05Customizable context injection for project-specific analysis

060 GitHub stars

Use Cases

01Validating complex system architectures against multiple AI perspectives to ensure robustness

02Conducting deep-dive technical audits and red-teaming of critical codebases

03Resolving ambiguous requirements by identifying points of divergence among leading LLMs

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add twwoo0210/claude-code-council-advanced council

For use in Claude.ai and ChatGPT

Key Features

01Simultaneous multi-model querying of Claude, GPT-5.3 Codex, and Gemini 3.1 Pro

02Automated synthesis highlighting consensus, divergence, and audited recommendations

034-Phase 'Team Deliberation' mode for structured internal reasoning

044-Round adversarial 'Deep Dive Debate' mode for high-stakes problem solving

05Customizable context injection for project-specific analysis

060 GitHub stars

Use Cases

01Validating complex system architectures against multiple AI perspectives to ensure robustness

02Conducting deep-dive technical audits and red-teaming of critical codebases

03Resolving ambiguous requirements by identifying points of divergence among leading LLMs

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add twwoo0210/claude-code-council-advanced council

For use in Claude.ai and ChatGPT