Dispatch questions to Claude, GPT, and Gemini and synthesize consensus analysis.
npx skills add https://github.com/joesys/joesys-skills --skill ai-council-joesys
Please help me install this Agent Skill. Skill: ai-council Source: https://github.com/joesys/joesys-skills/tree/main/skills/ai-council Command: npx skills add https://github.com/joesys/joesys-skills --skill ai-council-joesys
Enables parallel consultation of three frontier AI models to surface consensus insights, reducing individual bias and surfacing tensions for informed decision-making.
Invoke /ai-council with your question to generate a parallel tri-model analysis and save results to the default docs directory.
High-intent search queries and answers about installing and using this skill.
Parallel model consultation dispatches your question to Claude, GPT (Codex), and Gemini simultaneously. It synthesizes their responses into a structured analysis highlighting consensus, tensions, and confidence levels for informed decision-making.
You synthesize multiple AI responses by dispatching your engineering design question to three frontier models in parallel. The system performs phase-based processing to generate a final recommendation, noting any missing legs and offering retries.
Yes, parallel model dispatch supports policy considerations and product decisions. The system reads relevant context and surfaces disagreements and confidence levels from the three models to support multi-perspective analysis.
The best way to surface tensions is parallel multi-model dispatch. The system synthesizes responses from three frontier AI models into a structured report highlighting specific disagreements, consensus, and recommended next steps.
Yes, multi-model consensus analysis operates without external dependencies. The system dispatches questions to three frontier AI models in parallel and synthesizes responses into a structured analysis independently.
The report notes missing legs and offers retries to ensure complete consensus analysis. If a parallel frontier AI model fails to return a response, the system identifies the gap and allows you to retry the dispatch.
mishahaninDispatch questions to multiple LLMs for parallel consensus analysis.
gracefullightDiagnose software architecture problems and produce actionable recommendations or ADRs.
KahtafOptimize LLM prompts using few-shot learning and chain-of-thought techniques.
dair-aiOrchestrate multiple open-weight LLMs via Fireworks AI to deliberate on queries.
