Magi

Put a hard question to three independent LLM CLIs from different vendors and return a tallied verdict (decision mode) or a cross-checked research report (research mode), always with a matrix showing who said what. The three sages are MELCHIOR (Claude), BALTHASAR (Codex) and CASPER (Grok); they are configurable. Use it when one model's blind spot would be expensive: choosing between options, or checking whether a "current" fact is still true. English triggers: "ask the MAGI council", "put this to the MAGI", "three-model council", "get three independent opinions", "cross-check this with three models" 日本語トリガー: 「MAGI に諮って」「3賢者で合議」「3体で合議」「3モデルで多数決」「3つのモデルで裏取りして」

anyoneanderson Updated 3 repo stars

File contents

anyoneanderson/agent-skills commit cc40e79fea

Frequently asked questions

npx get-skill add anyoneanderson/magi