agent-browser

Automate browser navigation and UI actions via snapshot-ref interaction cycles.

4|1|Updated Mar 22, 2026
One-click install
npx skills add https://github.com/lgwanai/openforce --skill agent-browser-lgwanai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/lgwanai/openforce/tree/main/skills/agent-browser
Command: npx skills add https://github.com/lgwanai/openforce --skill agent-browser-lgwanai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill solves the problem of making AI agents reliably interact with real websites by providing a repeatable browser-automation workflow for navigation, element discovery, and UI actions.

Core Features & Use Cases

  • Browser interaction workflow: Open a page, take snapshots to get stable element references, perform clicks/fills/selections, then re-snapshot after each DOM change.
  • Web data capture & verification: Extract text, capture screenshots/PDFs, and use diffing (snapshot/screenshot) to confirm changes.
  • Operational robustness for automation: Supports chaining commands, session persistence, timeouts/explicit waits, domain allowlists, and action policies to reduce failure modes and unsafe behavior.
  • Use Case: You need to navigate to a production web app, log in, fill a multi-step form, take evidence screenshots, and extract structured results after each step.

Quick Start

Use the browser automation workflow to fill and verify a target form by running the agent-browser commands to open the page, snapshot interactive elements, fill fields using the returned refs, click submit, wait for the result, then snapshot again to confirm.

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 form filling and login flows with an AI agent?

Automate web form filling and login flows by using a snapshot-ref interaction cycle to discover elements, perform clicks and fills, then re-snapshot after DOM changes to verify progress.

How does browser automation handle dynamic DOM changes during web interaction?

Browser automation handles dynamic DOM changes by taking a snapshot to get stable element references, performing an action like a click or fill, and then taking a new snapshot to confirm the updated state.

Can I capture screenshots and extract structured data during web scraping?

Yes, you can capture screenshots and extract structured data during web scraping by applying built-in commands to snapshot text, take visual evidence, and generate PDFs of the live pages.

Does browser automation support session persistence for multi-step workflows?

Yes, browser automation supports session persistence, allowing you to chain commands across multi-step workflows, apply explicit waits, and maintain state during tasks like end-to-end verification.

What security controls exist for agent browser automation on production web apps?

Security controls for agent browser automation include domain allowlists and action policies, which restrict navigation and unsafe behavior to ensure operational robustness on production web apps.

How do I verify UI changes after an automated click on a web page?

Verify UI changes after an automated click by using snapshot or screenshot diffing to compare the page state before and after the action, confirming that the expected visual or DOM updates occurred.