What problem does it solve? Getting a second opinion on a design, code review, or security question usually means trusting a single model's answer. This Skill consults multiple AI models in parallel through OpenRouter, assigns each a professional role or stance, and produces a synthesized verdict with consensus points, disagreements, and actual cost reporting. ## Core Features & Use Cases - Tiered Review Mode: Builds a structured IT review team at levels 1-3 (3, 6, or 8 models) with domain-specific professional roles for code review, security, architecture, or general questions, with per-level cost caps from $0.50 to $10.00. - Flexible Panel Mode: Runs an ad-hoc panel of user-named models with custom stances (for/against) or roles from a catalog of 19 role definitions. - Cost Control and Failover: Estimates costs before running, enforces hard cost caps, retries failed calls, and substitutes fallback models within the same budget. - Use Case: Before merging a risky refactor, ask for a level 2 architecture review: six models acting as system architect, performance engineer, and other roles each critique the change, and you receive a synthesis of where they agree, where they disagree, and a recommendation. ## Quick Start Ask the assistant to run a tiered review of your design document at level 1, for example: "Run a level 1 panel review of this architecture proposal."
