What problem does it solve? High-stakes decisions suffer when a single perspective misses risks, and group discussion can collapse into groupthink. This Skill polls multiple expert personas as isolated parallel subagents so their independent answers reveal exactly where a decision is contested and where it is safe. ## Core Features & Use Cases - Blind Parallel Panel: Spawns 3-5 persona subagents in one batch via the Agent tool, each answering in isolation with no sight of the others. - Structured Verdicts: Forces every panelist to return the same schema (position, confidence, key reason, top risk, what would change my mind) so answers can be diffed mechanically. - Disagreement Map Synthesis: Aligns answers field-by-field to surface settled points, the axis of disagreement, lone dissenters, a confidence read, and the decisive question that would resolve the split. - Use Case: Before committing to a new architecture, convene Atlas, Forge, and Anchor blind; if Atlas flags scaling risk while Forge sees none, you know exactly which fact to verify before committing. ## Quick Start Ask the assistant to convene the council on whether your team should adopt a specific technology, naming the personas you want on the panel.