gstack

Automate browser-based QA testing with a persistent Playwright headless daemon.

Updated Apr 25, 2026
One-click install
npx skills add https://github.com/uzumaki-inc/uzustack --skill gstack-uzumaki-inc
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gstack
Source: https://github.com/uzumaki-inc/uzustack/tree/main/_upstream/gstack
Command: npx skills add https://github.com/uzumaki-inc/uzustack --skill gstack-uzumaki-inc

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires playwright, puppeteer-core, and includes scripts (resource) components.

What problem does it solve?

This powerful QA tool enables automated browser-based testing and site dogfooding by providing a fast, persistent way to drive a headless browser for QA, capture evidence, and verify state across pages.

Core Features & Use Cases

  • Navigate pages, interact with elements, and validate UI state.
  • Diff changes after actions and generate annotated screenshots to reproduce bugs.
  • Test multi-step flows, forms, file uploads, dialogs, and responsive layouts in real time.

Quick Start

Start a browse session in Claude Code and begin auditing a site using the /browse command.

Frequently Asked Questions about gstack

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

FAQPage Schema
How do I automate web testing and capture screenshots of UI defects?

Automated web testing with screenshot capture is performed by driving a persistent headless browser daemon to navigate pages, interact with elements, and reproduce defects. It validates UI state across multi-step flows using Playwright, generating annotated screenshots for evidence.

Can I test multi-step forms and dynamic content using Playwright in Claude Code?

Yes, multi-step forms and dynamic content can be tested using a Playwright-based daemon within Claude Code. The daemon maintains per-session state, allowing you to navigate complex workflows, handle file uploads, and interact with dialogs across multiple pages.

How does browser-based QA diffing work for reproducing bugs?

Browser-based QA diffing works by capturing UI state before and after an action, then generating annotated screenshots to highlight visual changes. This process reproduces bugs by providing visual evidence of the differences caused by specific interactions during testing.

Do I need Playwright or Puppeteer installed to run headless browser automation?

Yes, Playwright and puppeteer-core are required dependencies to run this headless browser automation. The daemon relies on these frameworks to execute browser-based QA testing, maintain per-session state, and capture screenshots of web pages.

What is the best way to dogfood a website and validate responsive layouts end-to-end?

The best way to dogfood a website and validate responsive layouts is by using a persistent headless daemon to drive real-time browser interactions. This approach tests multi-step flows and responsive layouts, capturing annotated screenshots to verify state and document defects.

Why does my headless browser testing lose state across multi-page flows?

Headless browser testing loses state across multi-page flows if the daemon does not maintain per-session persistence. This tool solves that by applying a persistent headless daemon with Playwright, ensuring per-session state is kept intact during end-to-end testing and site dogfooding.