What problem does it solve? When producing deliverables (documents, content, plans), creators tend to approve their own work too easily. This Skill acts as a strict review gate that prevents rubber-stamping by requiring evidence-based judgments before any output moves to the publishing stage. ## Core Features & Use Cases - Dual-axis review: Checks quality against predefined acceptance criteria (Axis A) and audits plan fulfillment item-by-item (Axis B), rating each as complete, partial, missing, done differently, or unverifiable. - Severity-rated findings: Classifies every finding as high, medium, or polish, then issues an overall verdict of pass, conditional pass, or fail with documented reasoning. - Parallel independent review: For large deliverables, spawns sub-agents from multiple angles (recipient perspective, criteria alignment, adversarial critique) and prioritizes findings confirmed by multiple angles. - Use Case: After drafting a multi-part content series, run this Skill to verify each piece meets the acceptance criteria, confirm the distribution path is connected, and route major gaps back to the planning or production stage. ## Quick Start Ask the AI to review the deliverable produced in the previous stage by saying "이거 됐나 냉정하게 검토해줘" (review this critically to see if it passes).
