What problem does it solve? Getting genuinely diverse perspectives from multiple AI agents is hard when a single model roleplays every voice, since opinions tend to converge and feel performative. This Skill solves that by spawning each BMAD agent as an independent subagent with its own persona, producing authentic disagreement and domain-specific insight. ## Core Features & Use Cases - Independent Subagent Spawning: Each selected agent runs as its own subagent process via the Agent tool, with persona built from the BMAD agent manifest (displayName, role, communication style, principles). - Flexible Orchestration: Pick 2-4 relevant voices per round, run targeted follow-ups (one agent reacting to another), and keep discussion context summarized under 400 words. - Solo and Model Control Modes: Use --solo to roleplay all agents in one response when subagents are unavailable, and --model to force a specific model for all subagents. - Use Case: A product team debating an architecture decision activates party mode, and the architect, developer, and QA agents each give independent takes, disagree on trade-offs, and react to each other's points across multiple rounds. ## Quick Start Start party mode and have the BMAD agents discuss the trade-offs of migrating our monolith to microservices.
