claude-council

A Claude Code plugin that consults multiple AI coding agents in parallel and shows you their answers side-by-side. Useful when one model's bias could mislead you and the right call depends on cross-checking — architecture decisions, debugging dead ends, security reviews, framework picks.

Five providers answering the same question. Each banner names the provider, the model that answered and how long it took. The synthesis separates what they all agreed on from where they diverged — here, whether trapping `EXIT INT TERM` on one handler is correct, or whether signals should be converted into exits first — and when they agree instead, it names the assumption the answer rests on.