gstack

Operate a persistent headless Chromium daemon for automated QA and regression checks.

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

SYSTEM DOCUMENTATION & REQUIREMENTS

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

What problem does it solve?

gstack provides a persistent, fast headless Chromium environment so agents can interact with real web pages for QA, dogfooding, and regression verification without paying browser startup or state-loss costs on every command.

Core Features & Use Cases

  • Persistent headless Chromium daemon with Bun-compiled CLI for sub-second follow-up commands.
  • Snapshot system that assigns accessibility-based @refs, supports annotated screenshots, unified diffs, and cursor-interactive (@c) refs for hard-to-find click targets.
  • Interaction primitives: click, fill, upload, dialog control, network/console capture, iframe support, viewport/responsive testing, and chainable multi-step flows.
  • Cookie import from real Chromium profiles, headed-mode with a Chrome side-panel for live observation and a sidebar agent, and state save/load for reproducible sessions.
  • Use cases: verify staging vs prod deployments, reproduce and capture bugs with annotated evidence, run end-to-end QA flows, and generate visual diffs for PRs.

Quick Start

Invoke the browse QA flow to test a URL and produce evidence, for example ask: "Run /qa on https://staging.example.com and produce annotated screenshots and a bug report".

Frequently Asked Questions about gstack

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

FAQPage Schema
How do I run automated QA testing with a headless browser without paying startup costs on every command?

Automated QA testing with a headless browser avoids startup costs by using a persistent local Bun/Playwright daemon, enabling reproducible, sub-second subsequent commands for navigating pages and interacting with elements.

Can I import real browser cookies into a Playwright session for end-to-end verification?

Yes, you can import real browser cookies into a Playwright session using built-in cookie decryption helpers, allowing you to authenticate and run multi-step chains against real user state for regression checks.

How do I capture annotated screenshots and unified diffs for staging vs prod deployments?

Capture annotated screenshots and unified diffs by running multi-step chains in a headless Chromium interface, assigning accessibility @ref locators and generating visual diffs to verify staging vs prod deployments.

Does this headless browser automation tool support iframe interaction and viewport testing?

Yes, this headless browser automation tool supports iframe interaction and viewport testing through interaction primitives like click, fill, dialog control, and responsive viewport adjustments within chainable multi-step flows.

What is the best way to reproduce and capture bugs with annotated evidence in Chromium?

The best way to reproduce and capture bugs is using a persistent Chromium environment that assigns accessibility-based @refs, captures network and console logs, and generates annotated screenshots with cursor-interactive overlays.

Why does my Playwright browser automation lose state between multi-step flows?

Playwright browser automation loses state when a new browser instance is launched per command; using a persistent daemon with state save/load capabilities maintains reproducible sessions across multi-step flows.