What problem does it solve? Visual prior-art surveys of UI conventions are easy to write and hard to trust: a tidy-looking artifact can hide unchecked citations, conflated zero results, or padded coverage. This Skill acts as an independent acceptance gate that judges a finished survey artifact against a single authoritative bar before it is approved. ## Core Features & Use Cases - Condition-based review: Walks 40 numbered conditions covering the vocabulary map, per-angle search output, extract records, and the convention register, quoting the exact artifact text that fails each one. - Citation spot-checking: Verifies that every cited corpus actually contains the convention claimed, catching plausible-looking but unsupported records. - Proportionality filter: Approves thin-but-honest results for narrow UIs and revises only on named, unrecorded gaps, preventing padding. - Use Case: After producing a UI-pattern vocabulary map and per-angle search output for a new product, run this reviewer to receive exactly one verdict line — VERDICT: approve or VERDICT: revise — with grounded findings naming their condition numbers. ## Quick Start Use reviewing-visual-prior-art-survey to judge the vocabulary map and search output in my survey folder and give me the verdict with findings.
