skills.rest
haidonglethqb
haidonglethqbCOMMUNITY
Explore all skills by this author →
GitHub

ck:agent-browser

Automate browser tasks via agent-browser CLI with snapshot and refs.

Updated Feb 27, 2026
Search Tags:#automation#testing#ai#cloud#browser#snapshot#refs
One-click install
npx skills add https://github.com/haidonglethqb/CloudSchool --skill ck-agent-browser-haidonglethqb
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: ck:agent-browser
Source: https://github.com/haidonglethqb/CloudSchool/tree/main/.qwen/skills/agent-browser
Command: npx skills add https://github.com/haidonglethqb/CloudSchool --skill ck-agent-browser-haidonglethqb

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Browser automation for AI agents often suffers from context drift and brittle workflows. This Skill provides a stable, context-efficient CLI that uses a snapshot + refs paradigm to coordinate long-running autonomous tasks across web pages.

Core Features & Use Cases

  • Context-efficient navigation and interaction using stable element refs
  • Support for long-running AI sessions, self-verifying loops, and cloud/browser testing (Browserbase)
  • Deterministic workflows with snapshotting, state persistence, and optional cloud providers

Quick Start

Install and run the agent-browser CLI, then open a URL and begin automating with snapshot, click, and fill commands.

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 browser tasks for AI agents without causing context drift?▼

Browser automation for AI agents uses a context-efficient snapshot and refs paradigm to coordinate long-running tasks. This approach minimizes state leakage by capturing stable element references, ensuring deterministic multi-step navigation and form interactions across web pages.

Does browser automation work with cloud browser testing platforms like Browserbase?▼

Yes, browser automation supports cloud browser testing with Browserbase. It enables long-running autonomous AI sessions and self-verifying workflows by leveraging named sessions and optional cloud providers to execute deterministic tasks remotely.

What is the best way to run self-verifying browser automation workflows in long-running sessions?▼

The best way to run self-verifying browser workflows is using a context-efficient CLI with snapshotting and state persistence. By utilizing open, snapshot, click, and fill commands, AI agents can execute deterministic tasks and validate state across multi-step navigation.

How do I start automating web pages using the snapshot and refs system?▼

To start automating web pages, install and run the agent-browser CLI. You open a target URL, generate a snapshot to capture stable element refs, and then use click and fill commands to interact with the page, ensuring reliable, context-efficient navigation.

Why does browser automation fail during long-running autonomous AI sessions?▼

Browser automation often fails in long-running AI sessions due to context drift and brittle workflows. Without a context-efficient snapshot system, state leakage accumulates, degrading reliability; using stable element refs mitigates this by maintaining deterministic navigation.

Can I use video-enabled debugging for browser automation workflows?▼

Yes, video-enabled debugging is supported within browser automation workflows. By utilizing snapshotting and state persistence alongside named sessions, you can review multi-step navigation and form interactions, maximizing reliability for autonomous AI tasks.

Related Skills

Explore More Software Engineering Skills→
haxlyshaxlys

agentcore

Run agent-browser sessions on AWS Bedrock with AWS credential resolution.

Community
Advanced
marcusrbrownmarcusrbrown
23

agent-browser

Automate browser interactions using Vercel's agent-browser CLI and ref-based selectors.

Community
Intermediate
patricio0312revpatricio0312rev
5

agent-orchestration-planner

Design multi-step agent workflows with tools, retry logic, and state management.

Community
Advanced
Thanh-aperoThanh-apero
1

agent-browser

Automate web browsing tasks using a snapshot and refs paradigm.

Community
Advanced