What problem does it solve? Getting genuinely diverse perspectives from an AI is hard because a single voice tends to converge on safe consensus. This Skill runs a lively roundtable where multiple distinct agent personas debate, clash, and build on each other, surfacing angles no single voice would reach. ## Core Features & Use Cases - Multi-persona roundtables: Runs installed BMAD agents or custom personas as distinct voices in one woven conversation, with support for inline voicing, independent subagents, or persistent agent teams. - Custom party authoring: Guides creation of custom personas and named groups (e.g., a Code Review Crew or an AI focus-group panel distilled from customer data), saved via TOML customization overrides. - Per-party memory: Keeps an append-only memlog per party so rooms remember past sessions, alliances, and callbacks across conversations. - Use Case: Run the built-in Code Review Crew party in subagent mode so a security engineer, an adversary, an edge-case hunter, a craftsman, and a pragmatist independently review your code and then argue about what actually matters. ## Quick Start Ask the AI to start party mode with the Code Review Crew to review your latest changes, optionally passing --mode subagent for independent reviews.
