Socratic

socratic is a model-invoked Clarification Engine for user-owned decisions. Use it when a clarification wrapper (clarify, project-clarify, decision-map) needs to track a conversation.

Core loop

1. Receive the latest answer as evidence.
2. Update current understanding.
3. Mark newly resolved decisions.
4. Recompute dependencies and the frontier: every meaningful user decision whose prerequisites are already settled.
5. Ask the complete actionable frontier as a round, not one question at a time.
6. Return internal state plus a conversational projection to the calling wrapper.