skills.rest
Orinks
OrinksCOMMUNITY
Explore all skills by this author →
GitHub

visual-verdict

Compare screenshots against reference images and return a strict JSON verdict.

24|2|Updated Mar 12, 2025
Search Tags:#ui-testing#design-review#reference-images#json-output#image-diff#visual-verdict
One-click install
npx skills add https://github.com/Orinks/AccessiWeather --skill visual-verdict-orinks
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: visual-verdict
Source: https://github.com/Orinks/AccessiWeather/tree/main/.codex/skills/visual-verdict
Command: npx skills add https://github.com/Orinks/AccessiWeather --skill visual-verdict-orinks

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This tool enables precise, automated assessment of UI visual fidelity by comparing a generated screenshot against reference images and producing a strict JSON verdict.

Core Features & Use Cases

  • Deterministic verdict with score, pass/revise/fail, category_match, and detailed differences and suggestions.
  • Supports multiple reference images to validate layouts across variants.
  • Returns concise guidance for next design iterations, including a clear reasoning field.

Quick Start

Provide the path to the generated_screenshot and one or more reference_images, optionally a category_hint, and run the visual-verdict evaluation to obtain a JSON result.

Frequently Asked Questions about visual-verdict

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

FAQPage Schema
How do I automate visual UI testing by comparing a generated screenshot against reference images?▼

Automate visual UI testing by submitting a generated screenshot and reference images to receive a deterministic JSON verdict containing a score, pass/revise/fail status, and detailed differences for design review.

What is a deterministic visual verdict for UI checks and how does it work?▼

A deterministic visual verdict is a strict JSON output evaluating UI fidelity across layouts, typography, and component styling by applying consistent comparison rules between a current output screenshot and provided reference images.

Can I use multiple reference images to validate different layout variants during a design review?▼

Yes, you can provide multiple reference images to validate layouts across variants. The evaluation compares the generated screenshot against them and returns category matches, differences, and suggestions.

What JSON schema is returned when performing an automated image diff for component styling?▼

The returned JSON schema includes score, verdict, category_match, differences, suggestions, and reasoning, providing concise guidance for next design iterations based on the image diff results.

Do I need to specify a category hint when running a UI discrepancy check?▼

No, a category hint is optional. You only need to provide the generated screenshot and reference images to run the evaluation, though a hint can help focus the visual QA results.

Related Skills

Explore More Design & Creative Skills→
gtpgg1013gtpgg1013

visual-verdict

Compare generated UI screenshots with reference images to produce a deterministic JSON verdict.

Community
Intermediate
Goblin1024Goblin1024
6

visual-verdict

Compare UI screenshots against reference images and output a JSON verdict.

Community
Intermediate
edonghyunedonghyun
1

compare

Compare baseline and current UI captures into a side-by-side visual report.

Community
Advanced
ClementATHClementATH

visual-verdict

Compare UI screenshots against reference images and output a deterministic JSON verdict.

Community
Advanced