hs:agent-browser

Automate Chrome and Chromium browsers via CDP with accessibility-tree snapshots and element refs.

Updated Jul 19, 2026
One-click install
npx skills add https://github.com/Dozyboy/VSF --skill hs-agent-browser-dozyboy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hs:agent-browser
Source: https://github.com/Dozyboy/VSF/tree/main/Day2_VSF/Demo1/harness/plugins/hs/disabled-skills/agent-browser
Command: npx skills add https://github.com/Dozyboy/VSF --skill hs-agent-browser-dozyboy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires agent-browser, and includes references (resource) components.

What problem does it solve? Manual browser testing, form filling, and scraping are slow and error-prone, and many automation tools require heavy Playwright or Puppeteer setups. This Skill provides a fast Rust-based CLI that drives Chrome/Chromium over the Chrome DevTools Protocol, letting an AI agent navigate pages, interact with elements, capture screenshots, and run QA flows without a real user profile. ## Core Features & Use Cases - Browser automation via CDP: Open pages, take accessibility-tree snapshots with stable @eN element refs, click, fill forms, and capture screenshots across sessions. - Specialized workflows: Load on-demand skill content for Electron apps, Slack automation, exploratory QA/bug hunts, and cloud browsers (Browserbase, AWS AgentCore, Vercel Sandbox). - Use Case: You need to verify a new build in a real browser. Install agent-browser, run agent-browser skills get core to load current workflows, then open your staging URL, snapshot the page, fill and submit the login form, and screenshot the result. ## Quick Start Ask the agent to install agent-browser with npm, load its core workflow content, then open a URL and take an interactive snapshot of the page.

Frequently Asked Questions about hs:agent-browser

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

FAQPage Schema
How do I automate browser testing with agent-browser?▼

Install the CLI with npm install -g agent-browser, run agent-browser install to download Chromium, then load workflows with agent-browser skills get core. Open a URL, run agent-browser snapshot -i to get element refs, and interact using commands like click @e1.

What is the difference between agent-browser and chrome-profile?▼

agent-browser uses a fresh, tool-managed browser for testing, scraping, and automation. chrome-profile is required when the task needs the user's real Chrome profile, existing cookies, logged-in sessions, or a specific Google account.

Can agent-browser run on cloud browsers like Browserbase?▼

Yes. Set BROWSERBASE_API_KEY and BROWSERBASE_PROJECT_ID environment variables, then run agent-browser -p browserbase open <url>. It also supports AWS Bedrock AgentCore and Vercel Sandbox via dedicated skill content.

Does agent-browser support Electron apps like VS Code or Slack?▼

Yes. Run agent-browser skills get electron to load workflows for Electron apps such as VS Code, Slack, Discord, Figma, Notion, and Spotify. A separate slack skill covers Slack workspace automation.

Why does agent-browser report element not found errors?▼

Element refs become stale after page changes. Re-run agent-browser snapshot -i to refresh the accessibility tree and obtain updated @eN references before interacting again.