browser

Control a real Chrome browser via CDP to automate and inspect web pages.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/Bad-Robots/BrowserHarness --skill browser-bad-robots
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: browser
Source: https://github.com/Bad-Robots/BrowserHarness/tree/main
Command: npx skills add https://github.com/Bad-Robots/BrowserHarness --skill browser-bad-robots

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

It gives agents direct, editable control of a real Chrome browser when browser tasks need to happen in the user's live session instead of a sandboxed automation layer.

Core Features & Use Cases

  • Direct CDP control for navigation, tab switching, typing, scrolling, screenshots, dialogs, uploads, and raw browser commands.
  • Support for local GUI Chrome, remote CDP endpoints, cloud browsers, and managed headless sessions.
  • Practical for browser QA, scraping, UI validation, site-specific automation, and reusable helper workflows.

Quick Start

Ask the agent to connect to your running Chrome, open the target page, and print the current page info.

Frequently Asked Questions about browser

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

FAQPage Schema
How do I automate Chrome browser tasks using CDP?

Chrome automation via CDP provides direct control over a real browser session to navigate, type, scroll, and handle dialogs. It enables agents to interact with live web pages, take screenshots, and execute raw browser commands for scraping and testing workflows.

Does browser automation work with remote CDP endpoints and headless sessions?

Yes, browser automation works with remote CDP endpoints, cloud browsers, and managed headless sessions. It supports connecting to external browser instances to perform UI validation, scraping, and automated testing workflows outside of a local GUI environment.

How do I handle stale tabs and browser permission prompts during web scraping?

To handle stale tabs and browser permission prompts during web scraping, the automation workflow manages browser sessions and safely clears stale profiles. It intercepts and manages browser permission prompts to ensure uninterrupted interaction with target web pages.

What is the best way to capture screenshots and validate UI in a live Chrome session?

The best way to capture screenshots and validate UI in a live Chrome session is through screenshot-driven interaction using CDP. This approach controls a real local GUI Chrome browser to inspect web pages, verify visual elements, and automate site-specific validation workflows.

Can I use raw CDP commands for file uploads and tab switching?

Yes, you can use raw CDP commands for file uploads and tab switching. The automation skill provides direct CDP access to handle file uploads, switch between multiple tabs, manage dialogs, and execute raw browser commands for comprehensive web page interaction.