verify-in-browser

Verifies visible UI changes by driving the running app in a browser before reporting completion.

Updated Jan 12, 2026
One-click install
npx skills add https://github.com/brandonarbini/arbini.family --skill verify-in-browser-brandonarbini
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: verify-in-browser
Source: https://github.com/brandonarbini/arbini.family/tree/main/.agents/skills/verify-in-browser
Command: npx skills add https://github.com/brandonarbini/arbini.family --skill verify-in-browser-brandonarbini

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Green test runs can hide broken UI: in a repository whose test environment is Node with no DOM, no test renders a component, so a change can break the browser render while every test passes. This Skill enforces browser-based verification of any user-visible change before it is reported as done. ## Core Features & Use Cases - Browser-driven verification: Open the affected page with browser automation (Claude in Chrome, Playwright, or a screenshot tool), perform the action, and inspect the result. - State and variant coverage: Check empty, loading, error, and long-content states, plus both locales, both themes, mobile width, and a second tenant for tenant-scoped apps. - Email template preview: Verify templates under emails/ via the react-email preview (pnpm dev:email) instead of sending real messages. - Honest reporting: Report the page, action, and observed result, or explicitly mark the change as unverified when the app could not be run. - Use Case: After editing a settings page layout, load /settings at mobile width in the browser, confirm the header renders correctly in both locales, and attach a screenshot to the final report. ## Quick Start Ask the assistant to verify the recent UI change in the browser by opening the affected page, exercising the empty and error states, and reporting what it actually saw.

Frequently Asked Questions about verify-in-browser

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

FAQPage Schema
How do I verify a UI change when tests do not render components?▼

Drive the running app in a browser: start the dev server, open the affected page with browser automation or a screenshot tool, perform the action, and inspect the result. A green test run in a Node environment without a DOM says nothing about rendering.

What should I check in the browser before reporting a frontend change as done?▼

Check the happy path plus empty, loading, error, and long-content states, then every variant the repository has: both locales, both themes, mobile width, and a second tenant for tenant-scoped apps. Include a screenshot in the report.

Can I use Playwright to verify Next.js pages visually?▼

Yes. The Skill explicitly supports Playwright scripts, Claude in Chrome, or screenshot tools to open pages served by the Next.js dev server and capture visual evidence of the change.

How do I preview react-email templates without sending emails?▼

Run the react-email preview with `pnpm dev:email` and open the template there. The preview renders the same component the send path uses, so it is a faithful verification target.

Why do tests pass while the page breaks in the browser?▼

When the test environment is Node with no DOM, no test renders a component, so render-breaking changes never fail a test. Only driving the real app in a browser surfaces these failures.

What should I report if I cannot run the app to verify a change?▼

State that the change is unverified rather than letting a green test suite imply otherwise. Name what you could not check so reviewers know the UI evidence is missing.