skills.rest
The-Rabak
The-RabakCOMMUNITY
Explore all skills by this author →
GitHub

agent-browser

Automate browser interactions via Bash CLI with ref-based element selection.

Updated Mar 3, 2026
Search Tags:#automation#cli#agent#browser#web-scraping#ref-based
One-click install
npx skills add https://github.com/The-Rabak/naysmith --skill agent-browser-the-rabak
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/The-Rabak/naysmith/tree/main/.github/skills/agent-browser
Command: npx skills add https://github.com/The-Rabak/naysmith --skill agent-browser-the-rabak

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Enables AI agents to programmatically drive web pages, fill forms, take screenshots, and scrape data via a lightweight Bash CLI, eliminating manual browser interactions and enabling repeatable automation.

Core Features & Use Cases

  • Navigation and control of web pages using open, back, forward, reload, and close commands
  • Snapshot generation of the accessibility tree (full, interactive, JSON) for robust element references
  • Interaction primitives including click, fill, type, hover, check, uncheck, and select
  • Information retrieval commands for text, HTML, attributes, title, URL, and element counts
  • Screenshots and PDF export for evidence, verification, and reporting
  • Sessions and semantic locators for parallel workflows and alternative element targeting

Quick Start

Open a URL with agent-browser open, snapshot the interactive elements, and then interact using refs to perform the task.

Frequently Asked Questions about agent-browser

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

FAQPage Schema
How do I automate browser web scraping for AI agents using a CLI?▼

Browser web scraping for AI agents is automated using a Bash CLI with ref-based element selection to navigate pages, fill forms, click elements, and extract data across single or multi-page workflows. It provides commands for navigation, interaction, snapshots, and session management.

How do I select elements for browser automation without CSS selectors?▼

Element selection for browser automation uses ref-based targeting generated from accessibility tree snapshots, with optional semantic locators for robust targeting. You snapshot interactive elements to get references, then pass those refs to interaction commands like click, fill, or hover.

Can I take screenshots and export PDFs during web scraping workflows?▼

Screenshots and PDF exports are supported during web scraping workflows for evidence, verification, and reporting. These commands run alongside navigation and interaction primitives to capture page states at any point in the automation flow.

What is the best way to run parallel browser automation sessions?▼

Parallel browser automation workflows are supported through built-in session management commands. You can open multiple independent sessions and use semantic locators and ref-based snapshots to drive concurrent page interactions across different tabs or contexts.

Does this browser automation tool work with multi-page form filling flows?▼

Multi-page form filling flows are fully supported through navigation commands like open, back, forward, and reload combined with interaction primitives including fill, type, check, uncheck, and select. You can chain these commands across pages using session management.

How do I extract text and HTML attributes from web pages programmatically?▼

Text, HTML, attributes, titles, URLs, and element counts are extracted programmatically using dedicated information retrieval commands. These commands query the current page state by leveraging the accessibility tree snapshots to return targeted data.

Related Skills

Explore More Software Engineering Skills→
lackeyjblackeyjb
3.0k

Playwright Browser Automation

Automate browser testing and interactions with Playwright scripts.

Community
Intermediate
firecrawlfirecrawl
121

firecrawl-workflows

Automate Firecrawl web workflows to generate structured reports and data extracts.

Official
Advanced
ProyectoG007ProyectoG007

agent-browser

Automate headless browser workflows using accessibility tree snapshots for element selection.

Community
Advanced
linaproailinaproai
147

playwright-cli

Automate browser interactions and generate Playwright test code via CLI.

Official
Intermediate