What problem does it solve? Getting genuinely independent perspectives from an AI is hard — a single model voicing multiple viewpoints tends to blend them into one agreeable answer. This Skill runs a moderated roundtable where distinct personas (installed BMAD agents or custom characters) argue, clash, and build on each other, optionally backed by real independent subagents so each voice actually thinks for itself. ## Core Features & Use Cases - Multi-persona roundtables: Run lively group discussions with installed agents or custom personas, with distinct voices, conflict, and running dynamics across rounds. - Four execution modes: Choose session (inline voicing), auto, subagent (a real agent per persona), or agent-team (persistent team, Claude Code only) depending on how independent the thinking needs to be. - Party authoring and memory: Create custom personas and named groups from ideas or raw customer data (e.g., an AI focus-group panel), with per-party persistent memory across sessions. - Use Case: Point the built-in Code Review Crew at a pull request — a security engineer, an adversary, an edge-case hunter, a craftsman, and a pragmatist each attack the code from their own lens and argue about what actually matters before it ships. ## Quick Start Ask the AI to start party mode so the installed agents can debate your topic, or request a specific room with a command like "run the code-review-crew party on this diff in subagent mode".
