gstack

Automate browser QA sessions with a persistent Playwright-driven daemon.

1|Updated Apr 26, 2026
One-click install
npx skills add https://github.com/tblakex01/conductor-playground --skill gstack-tblakex01
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: gstack
Source: https://github.com/tblakex01/conductor-playground/tree/main/.claude/skills/gstack
Command: npx skills add https://github.com/tblakex01/conductor-playground --skill gstack-tblakex01

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Gstack provides a persistent, low-latency browser environment to QA test, dogfood, and debug Claude Code-powered workflows without reinitializing a browser for every action.

Core Features & Use Cases

  • Persistent Playwright-driven daemon for rapid, multi-page QA sessions.
  • Ref-based snapshot system for robust element targeting and reliable automation.
  • Batch endpoints to execute multiple commands in a single request and reduce latency.
  • Annotated screenshots and diffing to verify state changes between steps.
  • Local-first design with per-session tokens and activity logging for auditability.

Quick Start

Launch the gstack CLI to start the headless browser and begin a QA session.

Frequently Asked Questions about gstack

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

FAQPage Schema
How do I automate browser-based QA testing without reinitializing the browser for every action?

You can use a persistent Playwright-driven daemon to maintain a low-latency browser environment for multi-page QA sessions. This approach keeps the browser active between actions, enabling rapid automated testing, design review, and debugging without repeated browser reinitialization overhead.

What is ref-based element targeting for browser automation and how does it work?

Ref-based element targeting uses a snapshot system to robustly identify and interact with page elements during automated browser workflows. By referencing structural snapshots, the automation maintains reliable element detection across multi-page workflows even when page content dynamically changes.

How do I execute multiple browser commands in a single request to reduce automation latency?

Batch endpoints allow you to execute multiple browser commands in a single request, significantly reducing latency. By grouping automation actions into one batch call, the persistent browser daemon processes them sequentially without the overhead of individual network round trips.

Can I test cross-tab workflows across local and remote pages using Playwright automation?

Yes, cross-tab testing across local and remote pages is supported by the persistent Playwright-driven daemon. It manages multi-page workflows with per-session isolation, ensuring that separate local and remote page contexts remain distinct while allowing synchronized automated QA actions.

Does headless browser QA automation require token-based authentication for session isolation?

Yes, token-based authentication is required to enforce per-session isolation in the headless browser environment. This local-first design ensures that individual QA sessions remain secure and isolated, while activity logging provides a complete audit trail for debugging and verification.

How do I verify state changes between steps in an automated browser workflow?

Annotated screenshots and diffing are used to verify state changes between steps in an automated browser workflow. The persistent daemon captures annotated visual snapshots at each stage, allowing you to compare page states and confirm that design or functional changes occurred as expected.