browse

Controls a persistent headless Chromium browser for QA testing and site verification.

Updated Jan 13, 2026
One-click install
npx skills add https://github.com/psyche911/AntigravityTest --skill browse-psyche911
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: browse
Source: https://github.com/psyche911/AntigravityTest/tree/main/.claude/skills/gstack/browse
Command: npx skills add https://github.com/psyche911/AntigravityTest --skill browse-psyche911

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires playwright, diff.

What problem does it solve? Manual QA testing and deployment verification require opening a browser, clicking through flows, and gathering evidence by hand. This Skill gives an AI agent a fast persistent headless Chromium session to navigate pages, interact with elements, and capture proof automatically. ## Core Features & Use Cases - Interactive Page Testing: Navigate URLs, click elements, fill forms, upload files, and handle dialogs using CSS selectors or snapshot refs like @e3. - State Verification & Diffing: Assert element states (visible, enabled, checked), diff snapshots before and after actions, and compare two environments such as staging versus production. - Evidence Capture: Take annotated screenshots, responsive layout screenshots across mobile/tablet/desktop viewports, and collect console errors and network request logs. - Session Reuse: Import cookies from real Chromium browsers (Chrome, Arc, Brave, Edge, Comet) via a local cookie picker UI to test authenticated flows. - Use Case: After deploying a new login flow, run snapshot to list interactive elements, fill credentials, submit, then diff the snapshot and screenshot the dashboard to confirm the flow works. ## Quick Start Ask the agent to open your app's URL with the browse tool, fill in the login form, submit it, and verify the dashboard appears with a screenshot.

Frequently Asked Questions about browse

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?▼

Navigate to the login page with goto, run snapshot -i to list interactive elements, then fill the username and password fields using their @refs and click submit. Verify success with is visible on a dashboard selector or snapshot -D to diff the page change.

How to take screenshots for a bug report using Playwright?▼

Run snapshot -i -a -o /tmp/annotated.png to capture a screenshot with labeled interactive element refs, or screenshot /tmp/bug.png for a plain full-page capture. Combine with console and network commands to attach error logs as evidence.

Can I import cookies from Chrome into a headless browser session?▼

Yes, the cookie-import-browser command reads and decrypts cookies from installed Chromium browsers on macOS including Chrome, Arc, Brave, Edge, and Comet. A local cookie picker UI at /cookie-picker lets you select domains to import into the Playwright session.

Does the headless browser session persist between commands?▼

Yes, a persistent server keeps Chromium alive between CLI calls, preserving cookies, tabs, and login sessions. The first call starts the server in about 3 seconds, subsequent commands run in roughly 100ms, and the server shuts down after 30 minutes idle.

Why did my click command fail with an element not found error?▼

Clicks fail when the selector matches nothing, matches multiple elements, or refs went stale after navigation. Run snapshot again to get fresh @refs, or use a more specific CSS selector before retrying the interaction.

How do I compare staging and production versions of a page?▼

Use the diff command with two URLs, for example diff https://staging.app.com https://prod.app.com. It loads both pages, extracts clean text, and outputs a unified line-by-line diff of the content differences.