visual-verdict

Compare UI screenshots against reference images and return a deterministic JSON visual QA verdict.

Updated May 17, 2026
One-click install
npx skills add https://github.com/tiankong0101-byte/skills-registry --skill visual-verdict-tiankong0101-byte
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: visual-verdict
Source: https://github.com/tiankong0101-byte/skills-registry/tree/main/skills/oh-my-claudecode/skills/visual-verdict
Command: npx skills add https://github.com/tiankong0101-byte/skills-registry --skill visual-verdict-tiankong0101-byte

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill removes ambiguity from visual QA by comparing a generated screenshot against reference images and returning a deterministic JSON verdict you can use to decide whether another edit pass is needed.

Core Features & Use Cases

  • Screenshot Comparison: Evaluates layout, spacing, typography, colors, and hierarchy against one or more reference images.
  • Structured Verdict Output: Produces a strict JSON response with a score, pass/revise/fail verdict, mismatch notes, and actionable suggestions.
  • Iterative Visual Review: Ideal for UI polish workflows such as dashboard recreation, social feed layouts, or any design task where pixel-level fidelity matters.

Quick Start

Ask for a visual verdict on your generated screenshot using the provided reference images and request a strict JSON result with improvement suggestions.

Frequently Asked Questions about visual-verdict

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I get a pass or fail verdict from a UI screenshot comparison?

You can automate visual QA by comparing a generated UI screenshot against reference images to receive a deterministic JSON verdict, including a score, pass/revise/fail status, and actionable suggestions.

What is visual QA and how does pixel diff work for design fidelity?

Visual QA evaluates design fidelity by checking layout, spacing, typography, colors, and hierarchy against reference images. The pixel diff identifies concrete differences and returns mismatch notes to guide iterative UI polish.

How do I run a visual review on a dashboard recreation screenshot?

Provide your dashboard recreation screenshot and reference images. The visual review checks component styling and hierarchy, then returns a structured JSON verdict with mismatch notes and improvement suggestions for your layout.

Can I compare a generated UI against multiple reference images for layout checks?

Yes, you can evaluate a generated UI screenshot against one or more reference images. The comparison checks layout, spacing, and component styling, returning a category match flag and concrete differences in the JSON verdict.

Does the visual verdict output structured JSON for automated UI review workflows?

Yes, the visual verdict produces a strict JSON response containing a score, verdict, category match flag, and brief reasoning. This deterministic output is designed to help you decide whether another automated edit pass is needed.

What are the limitations of using a JSON verdict for screenshot comparison?

The JSON verdict provides deterministic pass/revise/fail decisions based on visual QA checks for layout and styling. It focuses strictly on screenshot-to-reference review and does not modify the UI or generate new design assets.