council

Multi-agent debate improves risk management by coordinating four specialized agents and two cross-examiners to analyze tasks independently. This structured deliberation surfaces weaknesses and synthesizes a final, audit-friendly recommendation for complex decisions.

Multi-agent synthesis is a deliberation process where four specialist agents analyze a task independently, and two cross-examiners challenge their conclusions. This phase-based governance produces a final synthesized recommendation with an audit-friendly trail.

You need phase-based governance and role separation when a high-stakes outcome requires robust reasoning. It applies to complex tasks requiring diverse viewpoints, cross-examination, and structured disagreement to ensure higher-quality recommendations.

You start a structured deliberation by invoking the /council command. This triggers the orchestration of four specialist agents and two cross-examiners to begin analyzing your task and producing a synthesized recommendation.

Multi-agent deliberation works effectively for cross-domain tasks by applying four specialized agents with distinct thinking styles. It enables structured disagreement and synthesis across domains to produce reproducible, audit-friendly deliberation trails.

Multi-agent deliberation is not suited for everyday decisions due to the overhead of coordinating four specialist agents and two cross-examiners. It should be reserved for high-stakes outcomes and complex tasks requiring robust reasoning, diverse viewpoints, and adversarial analysis.