Adversarial multi-agent debate for PR reviews, architectural reviews, and feature reviews. Committed opposing agents attack each other's cases across bounded rounds; only evidence-backed conclusions survive to a single ruthless verdict. Use when the user invokes $adversarial-review, asks for an adversarial, red-team, or debate review, or wants a PR, architecture, or feature judged brutally on evidence, best practices, and current data without nitpicking.
