clarify

Pre-debate clarification ensures shared problem understanding by collecting and relaying agents' questions before a debate starts. The orchestrator coordinates a per-agent clarification loop, recording user answers to align all participants on the debated problem.

A pre-debate clarification loop is a multi-agent workflow process that collects questions from each configured agent iteratively. It records the user's answers for each question and persists them to the debate workspace to ensure alignment and auditability before the main debate begins.

To set up agent clarifications, enable clarifications in the debate config and run the orchestrator. The orchestrator then dispatches role-clarify subagents to start collecting per-agent questions, allowing multiple iterations per agent until all clarifications are resolved.

Yes, the multi-agent clarification process allows multiple iterations per agent. The orchestrator coordinates the loop across all configured agents, dispatching role-clarify subagents and recording the user's answers for each question until the clarification phase is complete.

Yes, the clarification orchestrator persists questions and answers to the debate workspace for auditability. It writes each agent's questions and the user's answers to the workspace, generating a final clarification summary document to reference during the debate.

The clarification orchestrator is limited to pre-debate alignment and does not participate in the actual debate. It requires configured agents and a debate workspace to function, meaning it cannot operate standalone without an existing multi-agent debate setup.