LLM Council: When

Purpose

Answer the "should I even do this?" question before anyone designs a pipeline, picks members, or writes a synthesizer prompt. Most tasks do not need a council. The default is a single strong model with a long thinking budget; a council has to earn its extra cost, not just be assumed better.

When to use / when NOT to use

Use this skill when the user is deciding whether to run multiple models and merge their outputs at all — "should I", "is it worth it", "council vs single model" — including when they describe the idea without naming it ("would 3 AIs beat 1 good one?").

Not for: