agent-browser-cli-control

Control Chrome tabs, execute JavaScript, and extract page content via agent-browser-cli.

4|Updated May 16, 2026
One-click install
npx skills add https://github.com/reason-machines/devtools-skills --skill agent-browser-cli-control-reason-machines
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-browser-cli-control
Source: https://github.com/reason-machines/devtools-skills/tree/main/skills/agent-browser-cli-control
Command: npx skills add https://github.com/reason-machines/devtools-skills --skill agent-browser-cli-control-reason-machines

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires @sleepinsummer/agent-browser-cli.

What problem does it solve? AI agents often need to interact with web pages that require the user's existing login sessions and cookies, which headless tools like Selenium or Playwright cannot access. This Skill lets agents control a real Chrome session through agent-browser-cli, preserving authentication state while automating browser tasks. ## Core Features & Use Cases - Tab and Page Automation: List, switch, open, and close Chrome tabs, then scan pages for simplified HTML, text-only content, or full DOM. - JavaScript Execution and CDP Operations: Run arbitrary JavaScript in any tab with optional DOM change monitoring, capture screenshots, and manage cookies per domain. - Use Case: An agent needs to extract data from an authenticated dashboard. It opens the URL in the user's Chrome, reads cookies for the domain, extracts the page content with a text-only scan, and saves a screenshot for verification. ## Quick Start Ask the agent to open a website in Chrome with agent-browser-cli and extract the page's text content.

Frequently Asked Questions about agent-browser-cli-control

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

FAQPage Schema
How do I control Chrome from the command line with an existing session?▼

Use agent-browser-cli, which connects to your running Chrome via a companion extension instead of launching a headless instance. After starting the daemon, commands like tabs, open, scan, and exec let you manage tabs and run JavaScript while keeping your cookies and login state.

agent-browser-cli vs Selenium or Playwright for browser automation?▼

agent-browser-cli attaches to an existing Chrome session through a Chrome extension, preserving all user logins and cookies. Selenium and Playwright typically launch fresh or headless browser instances, which lose authenticated sessions and require separate login handling.

How do I execute JavaScript in a browser tab from a CLI?▼

Run agent-browser-cli exec followed by your JavaScript code, and it returns the result as JSON. Add the --monitor flag to detect DOM changes after execution, or --tab to target a specific tab by index.

Why is the agent-browser-cli extension not connecting?▼

Verify the extension is loaded in chrome://extensions with Developer mode enabled, and confirm at least one normal web page tab is open rather than about:blank or chrome:// pages. Check the port in ~/.agent-browser-cli/config.json matches, then restart the daemon.

Does agent-browser-cli work on WSL 2?▼

Yes, but WSL 2 on Windows 11 22H2+ requires networkingMode=mirrored in .wslconfig so the Chrome extension can reach localhost:18765. Restart WSL after changing the setting and verify the extension connects.

What are the limitations of agent-browser-cli?▼

It requires Chrome running with the bridge extension loaded and at least one normal web page open, so it cannot run fully headless. Screenshot and monitoring operations are slower than text-only scans, and port conflicts on 18765 or 18767 require manual reconfiguration.