What problem does it solve? Single-perspective analysis carries blind spots, and decisions agreed to without friction go untested. This Skill stages a structured adversarial debate between two opposing AI personas over a plan, design, claim, or decision, surfacing weaknesses before you commit. ## Core Features & Use Cases - Structured Multi-Round Debate: Runs Opening, Rebuttal, and optional Counter rounds with parallel independent subagents, ending in a neutral synthesis verdict. - Automatic Persona Pairing: Auto-selects the natural opposing pair for your artifact type (e.g., scale-vs-simplicity for architecture, rigor-vs-feasibility for experiments), or uses personas you name. - Steelman Enforcement: Each side must present the opponent's strongest case before attacking it, preventing strawman arguments and forced harmony. - Use Case: Before committing to a microservices migration plan, run a debate between a scalability-focused persona and a ship-it pragmatist persona to expose hidden risks and get a clear call with a flip condition. ## Quick Start Ask the AI to debate your plan or decision, for example: "Debate this architecture proposal — should we migrate to microservices or keep the monolith?"