ck:agent-browser

Automate browser navigation, inspection, and interaction workflows with snapshot-driven refs.

Updated May 10, 2026
One-click install
npx skills add https://github.com/dthuy62/kollet-app --skill ck-agent-browser-dthuy62
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ck:agent-browser
Source: https://github.com/dthuy62/kollet-app/tree/main/.opencode/skills/agent-browser
Command: npx skills add https://github.com/dthuy62/kollet-app --skill ck-agent-browser-dthuy62

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

It removes the friction of manually driving web browsers by letting an AI open pages, inspect interactive elements, and complete actions with compact context.

Core Features & Use Cases

  • Snapshot-driven automation: Capture interactive elements, then act on stable refs instead of brittle selectors.
  • Long-running browser workflows: Handle navigation, forms, waits, tabs, screenshots, PDFs, and recordings across extended sessions.
  • Cloud and local testing: Run the same workflow in a local browser or on Browserbase for CI/CD and headless environments.
  • Use case: Use it to log into a site, verify a checkout flow, record the session, and save the authenticated state for later reuse.

Quick Start

Ask the AI to open your target website, take an interactive snapshot, and continue the task by clicking and filling the referenced elements.

Frequently Asked Questions about ck:agent-browser

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

FAQPage Schema
How do I automate web testing without brittle CSS selectors?

Web testing automation relies on snapshot-driven refs to capture interactive elements, allowing you to act on stable references instead of brittle selectors. This approach maintains reliable browser navigation and inspection throughout extended autonomous sessions.

Can I run browser automation workflows in a CI/CD environment?

Browser automation workflows support cloud execution with Browserbase for CI/CD and headless environments. You can run the same navigation, form interaction, and visual debugging tasks across both local browsers and cloud testing infrastructure.

How do I maintain authenticated state for long-running browser workflows?

Long-running browser workflows support session state persistence to save authenticated logins for later reuse. You can navigate forms, handle multi-tab testing, and resume extended sessions without repeatedly logging into target websites.

What is the best way to visually debug multi-tab web testing?

Visual debugging for multi-tab web testing uses screenshots, session recordings, and structured commands to verify browser tasks. You capture interactive snapshots across multiple tabs to inspect and confirm workflow completion efficiently.

Does browser automation support generating PDFs and recordings during a session?

Browser automation supports generating PDFs and recordings during active sessions. You handle navigation, forms, waits, and multi-tab interactions while capturing screenshots and recording the entire workflow for verification and debugging purposes.

Why does web testing fail when interactive elements load dynamically?

Web testing handles dynamically loaded interactive elements using structured waits and snapshot-driven refs. By capturing stable references after elements load, the automation avoids timing failures and maintains reliable interaction during long autonomous workflows.