llm-council-members

Purpose

Decide who sits on an LLM council and what role each seat plays. The single failure mode this skill exists to prevent: mistaking persona count for model diversity — five prompt variations on one model producing theatrical diversity (different voices) instead of epistemic diversity (different blind spots that can actually catch each other's errors). Deeper evidence, full citations, and persona-engineering detail live in references/composition.md.

When to use / when NOT to use

Use this skill when the user is choosing which models to seat on a council or what role/lens each seat should carry — including when they describe the problem without saying "council" (e.g. "is GPT+Claude+Gemini actually more diverse than three GPT personas?", "does it matter if my reviewers are secretly the same model in different hats?").