skills.rest
mikkelkrogsholm
mikkelkrogsholmCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

agent-browser

Automate browser interactions and extract structured web content via CLI daemon.

66|8|Updated Mar 13, 2026
Search Tags:#screenshots#cdp#session-management#web-scraping#browser-automation#form-filling#device-emulation
One-click install
npx skills add https://github.com/mikkelkrogsholm/skills --skill agent-browser-mikkelkrogsholm
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/mikkelkrogsholm/skills/tree/main/.agents/skills/agent-browser
Command: npx skills add https://github.com/mikkelkrogsholm/skills --skill agent-browser-mikkelkrogsholm

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

Automates repetitive and error-prone browser tasks so agents and scripts can reliably navigate websites, interact with UI elements, handle authentication, capture screenshots, and extract structured content without manual browsing.

Core Features & Use Cases

  • Snapshot-based refs for compact element targeting (e.g., @e1) that reduce context size for LLM-driven workflows.
  • Full interaction surface: open, wait, click, fill, select, upload, download, file/pdfs, screenshots, PDF export, and visual diffs for regression testing.
  • Session and state management options: import browser auth, persistent profiles, named sessions, encrypted state, and an auth vault for repeated authenticated automation.
  • Advanced capabilities: CDP connection, device emulation, proxy support, video recording, profiler traces, and action policies for safer automation.
  • Use case examples: automated login and data extraction for dashboards, end-to-end UI testing with visual diffs, scheduled scraping jobs with session reuse.

Quick Start

Open https://example.com, snapshot interactive elements, fill the login fields using a saved session or auth vault, submit and wait for the dashboard, then save a full-page screenshot and extract the main content as text.

Frequently Asked Questions about agent-browser

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

FAQPage Schema
How do I automate web scraping and browser interactions without manual browsing?▼

Browser automation automates web scraping and UI interactions via a CLI-driven daemon, enabling agents to navigate, fill forms, and extract content without manual browsing. It handles authentication, screenshots, and downloads reliably for testing or scraping workflows.

Can I reuse authenticated browser sessions for repeated scraping jobs?▼

Yes, authenticated browser sessions can be reused for repeated scraping jobs using persistent profiles, named sessions, and an auth vault. This enables importing browser auth and maintaining encrypted state for reliable repeated automated access.

What's the best way to capture screenshots and visual diffs for regression testing?▼

Capturing screenshots and visual diffs for regression testing is best handled through the browser daemon's built-in screenshot and visual diff capabilities. It supports full-page screenshots, PDF export, and visual diffing for end-to-end UI testing.

Does browser automation support device emulation and CDP connectivity?▼

Yes, browser automation supports device emulation and CDP connectivity. Advanced capabilities include Chrome DevTools Protocol connections, device emulation, proxy support, video recording, and profiler traces for configurable and safer automation workflows.

How do I handle form filling and file downloads in an automated browser workflow?▼

Form filling and file downloads are handled through the daemon's full interaction surface. You can fill, select, upload, download files or PDFs, and apply configurable waits and action policies to ensure reliable scripted browser interactions.

Do I need snapshot refs to reduce context size for LLM-driven browser workflows?▼

Yes, snapshot refs reduce context size for LLM-driven browser workflows by generating compact element targeting references like @e1. This allows agents to interact with web page elements efficiently without processing excessive page context.

Related Skills

Explore More Software Engineering Skills→
binggandatabinggandata
559

bggg-tiktok-search

Search TikTok videos and extract structured data via local Chrome session.

Community
Advanced
Pans0020Pans0020

markdown-exporter

Convert Markdown content into DOCX, PPTX, XLSX, PDF, HTML, IPYNB, MD, CSV, JSON/JSONL, XML, and extracted code files.

Community
Advanced
yasinnertenyasinnerten

test-scraper

Test HTML parsing and data extraction logic against diverse HTML structures.

Community
Basic
spideynolovespideynolove
2

lightpanda

Fetch HTML content from web pages using a lightweight headless browser.

Community
Basic