gstack

Automates QA testing and site verification through a persistent headless Chromium browser CLI.

Updated Mar 22, 2026
One-click install
npx skills add https://github.com/fullya99/simple-blog-page --skill gstack-fullya99
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: gstack
Source: https://github.com/fullya99/simple-blog-page/tree/main/.claude/skills/gstack
Command: npx skills add https://github.com/fullya99/simple-blog-page --skill gstack-fullya99

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires playwright, diff.

What problem does it solve? AI coding agents cannot see or interact with live web applications, forcing developers to manually click through pages, check console errors, and verify deployments by hand. This Skill gives the agent a fast headless browser so it can test user flows, verify deployments, and file bugs with screenshot evidence. ## Core Features & Use Cases - Ref-based page interaction: Snapshot the accessibility tree to get stable element refs, then click, fill forms, upload files, and handle dialogs without guessing CSS selectors. - Verification and evidence capture: Diff snapshots before and after actions, take annotated screenshots, check console and network logs, and assert element states like visible, enabled, or checked. - Authenticated testing: Import cookies from real browsers (Chrome, Arc, Brave, Edge) or JSON files to test logged-in pages without manual sign-in. - Use Case: After pushing a branch, ask the agent to verify staging — it logs in, walks the signup flow, screenshots each page, checks the console for errors, and reports a full QA pass in about a minute. ## Quick Start Ask the agent to use the gstack browse skill to open your staging URL, log in with test credentials, click through the main pages, and report any console errors with screenshots.

Frequently Asked Questions about gstack

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

FAQPage Schema
How do I test a login flow with a headless browser CLI?▼

Navigate with goto, run snapshot -i to list interactive elements with refs, then fill the username and password fields by ref and click submit. Verify success with snapshot -D to diff the page change and is visible to assert the dashboard appeared.

How to verify a deployment after pushing code?▼

Point the browser at your staging URL, then run text to confirm the page loads, console to check for JavaScript errors, network for failed requests, and screenshot for visual evidence. The whole check takes seconds since the browser stays running between commands.

Can I test authenticated pages without logging in manually?▼

Yes. The cookie-import-browser command decrypts and imports cookies from real Chromium browsers like Chrome, Arc, Brave, or Edge, either through an interactive picker UI or directly with a --domain flag. JSON cookie files are also supported.

Does the headless browser work on Linux servers and WSL?▼

Yes. The compiled binary and headless Chromium run on macOS, Linux, and WSL. Browser-based cookie decryption is macOS-only, but JSON cookie import works on all platforms, and the installer can add required system dependencies on Linux.

Why use a CLI instead of Playwright MCP for browser automation?▼

MCP calls carry JSON schemas and protocol framing that consume thousands of context tokens per call, while this CLI prints plain text to stdout with zero context overhead. Subsequent commands also run in 100-200ms versus 1-5 seconds for MCP tools.

What are the limitations of ref-based element selection?▼

Refs are invalidated on every navigation, so you must run snapshot again after goto. Elements missing from the accessibility tree, like clickable divs, require the snapshot -C flag to be discovered as cursor-interactive refs.