visual-verdict

Compare UI screenshots to reference images and output a strict JSON verdict.

Updated Apr 24, 2026
One-click install
npx skills add https://github.com/RJJJJ/project2 --skill visual-verdict-rjjjj
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: visual-verdict
Source: https://github.com/RJJJJ/project2/tree/main/.codex/skills/visual-verdict
Command: npx skills add https://github.com/RJJJJ/project2 --skill visual-verdict-rjjjj

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill automates the evaluation of UI visual fidelity by comparing a generated screenshot to one or more reference images and delivering a strict JSON verdict for design iteration.

Core Features & Use Cases

  • Deterministic, machine-readable verdicts for UI comparisons.
  • Supports multiple reference images and optional category hints to guide evaluation.
  • Integrates with design-review pipelines to quantify visual differences and provide actionable feedback.

Quick Start

Provide your generated_screenshot and one or more reference_images to obtain a JSON verdict.

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 QA checks for a UI screenshot against a reference image?

You automate visual QA checks by comparing a generated UI screenshot against reference images to produce a strict JSON verdict. This deterministic pass/fail guidance quantifies visual differences for design iteration.

What is a deterministic visual verdict in UI regression checks?

A deterministic visual verdict is a strict JSON output containing fields like score, verdict, differences, suggestions, and reasoning. It delivers machine-readable pass/fail guidance for UI regression checks by comparing screenshots to reference images.

Can I compare a generated UI screenshot against multiple reference images?

Yes, you can compare a generated UI screenshot against multiple reference images. The visual QA process accepts an array of reference images alongside the generated screenshot and an optional category hint to guide the evaluation.

How do I get structured JSON feedback for UI design reviews?

You obtain structured JSON feedback for UI design reviews by providing a generated screenshot and reference images to the visual QA process. It outputs a strict JSON object with score, verdict, differences, and suggestions for actionable iteration.

What format does the visual QA verdict output for regression pipelines?

The visual QA verdict outputs a strict JSON format for regression pipelines. This deterministic output includes specific fields: score, verdict, category_match, differences, suggestions, and reasoning, enabling automated design evaluation.

When should I use category hints in UI visual validation?

Use optional category hints in UI visual validation when you need to guide the evaluation of a generated screenshot against reference images for specific design contexts. They help tailor the deterministic JSON verdict for regression checks and visual QA pipelines.