skills.rest
compozy
compozyOFFICIAL
Explore all skills by this author →
Website𝕏@compozyaiGitHub

agent-browser

Automate browser interactions, extract page data, and capture screenshots.

182|8|Updated Apr 4, 2026
Search Tags:#data extraction#browser automation#web testing#form filling#screenshot#accessibility tree#ref targeting
One-click install
npx skills add https://github.com/compozy/agh --skill agent-browser-compozy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/compozy/agh/tree/main/.agents/skills/agent-browser
Command: npx skills add https://github.com/compozy/agh --skill agent-browser-compozy

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

agent-browser removes the manual effort of navigating websites, filling forms, taking screenshots, and extracting page information by turning common web interactions into repeatable steps.

Core Features & Use Cases

  • Browser navigation & page control: Open URLs, go back/forward, reload, and close sessions reliably.
  • Snapshot-based element targeting: Capture an accessibility tree (optionally interactive elements) and use stable refs to interact with the exact elements you need.
  • Form interactions, extraction, and screenshots: Click, fill, check/uncheck, select options, wait for conditions, extract text/values, and generate screenshots for verification.
  • Use case: For example, test a login flow by opening the login page, snapshotting interactive elements, filling credentials, submitting, waiting for the dashboard URL/load state, then extracting confirmation text and saving a screenshot.

Quick Start

Use agent-browser to open a URL, snapshot interactive elements, and then fill and submit a form using the returned refs.

Frequently Asked Questions about agent-browser

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

FAQPage Schema
How do I automate web testing for dynamic pages that require navigation and DOM-aware element selection?▼

Automate web testing by navigating URLs, snapshotting interactive elements via accessibility trees, and using stable refs to deterministically click, fill, and wait for page states. This approach handles dynamic pages requiring DOM-aware selection and generates screenshots for UI verification.

How does accessibility tree snapshotting work for targeting interactive elements during browser automation?▼

Accessibility tree snapshotting captures interactive elements on a page and assigns them stable refs. These refs allow deterministic browser automation commands to target exact elements for clicks, fills, and selections without relying on brittle CSS selectors.

Can I extract page information and generate machine-readable outputs from web pages?▼

You can extract page information by pulling text and values from targeted elements. The automation supports optional JSON outputs, allowing you to parse extracted page data into machine-readable formats for downstream processing.

What is the best way to automate form filling and submit login credentials on a website?▼

Automate form filling by opening the URL, snapshotting interactive elements, using the returned refs to fill credentials, and submitting the form. You can then wait for the dashboard URL or load state and extract confirmation text for verification.

Does browser automation support waiting for specific page conditions and taking screenshots for UI verification?▼

Browser automation supports deterministic wait conditions for page loads and state changes. After interactions complete, you can capture screenshots to verify UI states and ensure dynamic content has fully loaded.

What are the limitations of ref-based targeting for web testing on highly dynamic single-page applications?▼

Ref-based targeting relies on accessibility tree snapshots, so elements must be present in the tree when snapshotted. Highly dynamic single-page applications may require re-snapshotting after DOM mutations to capture newly rendered interactive elements.

Related Skills

Explore More Software Engineering Skills→
liuaibin001liuaibin001

automating-browsers

Automate browser interactions for AI agents using the agent-browser CLI.

Community
Advanced
alibabaalibaba
10.5k

pdf-extractor

Extract text, tables, and metadata from PDF documents into JSON.

Official
Intermediate
stanleykao72stanleykao72

field-recorder

Record dimensions, quantities, and photos into Odoo Spreadsheet from LINE chats.

Community
Advanced
hheydaroffhheydaroff

browser-tools

Automate web browsing, DOM interaction, and page content extraction.

Community
Intermediate