skills.rest
33may
33mayCOMMUNITY
Explore all skills by this author →
GitHub

browser

Automates web browser navigation, interaction, and data extraction using AI-optimized accessibility snapshots.

1|1|Updated Nov 28, 2025
Search Tags:#browser automation#web testing#web scraping#form filling#agent-browser#mcp tools#accessibility snapshot
One-click install
npx skills add https://github.com/33may/robotics --skill browser-33may
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: browser
Source: https://github.com/33may/robotics/tree/main/humanoid/.claude/skills/browser
Command: npx skills add https://github.com/33may/robotics --skill browser-33may

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Manual web browsing for testing, scraping, and form automation is slow and error-prone, and raw DOM dumps overwhelm AI context windows. This Skill provides browser automation with compact accessibility-tree snapshots that reduce context usage by 93% through element references. ## Core Features & Use Cases - AI-Optimized Snapshots: Capture interactive elements only with stable refs (@e1, @e2) instead of full DOM output. - Full Interaction Set: Click, fill, type, hover, select, scroll, and wait for elements, text, URLs, or network states. - Multi-Session Support: Run isolated parallel browser sessions and share saved authentication state between them. - Use Case: Automate a login flow by opening the page, snapshotting interactive elements, filling credentials via refs, clicking submit, and waiting for the dashboard URL. ## Quick Start Ask the agent to open a website with the browser skill, take an interactive snapshot, and click or fill elements using the returned refs.

Frequently Asked Questions about browser

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

FAQPage Schema
How do I automate browser interactions with an AI agent?▼

Open a URL with agent-browser, run snapshot -i to get interactive elements with refs like @e2, then use click, fill, or type commands against those refs. Re-snapshot after any page change since refs become stale.

How to fill and submit a web form using browser automation?▼

Navigate to the form page, take an interactive snapshot, then use fill with each input's ref to enter values and click the submit button's ref. Confirm success with wait --text or wait --url for the confirmation state.

What selectors does agent-browser support?▼

It supports element refs from snapshots, standard CSS selectors, and semantic locators by role, label, or test ID. Refs are recommended because they are deterministic and tied to the current accessibility tree.

Can I run multiple browser sessions in parallel?▼

Yes, use the --session flag to create isolated sessions, and list them with session list. You can save authentication state from one session and load it into another to share login context.

Why use snapshots instead of raw DOM for browser automation?▼

Snapshots provide an accessibility tree with element refs that reduces context usage by 93% compared to full DOM output. The -i flag further limits output to interactive elements, keeping agent context small and interactions reliable.

Related Skills

Explore More Software Engineering Skills→
MantisWareMantisWare

import

Import workspace templates from local files, URLs, or GitHub repositories with collision handling and secret prompts.

Community
Advanced
Runchuan-BURunchuan-BU
401

agent-browser

Automate web browsing to navigate pages, click elements, fill forms, and extract data.

Community
Advanced
vanducngvanducng
5

browser-trace

Capture read-only DevTools Protocol traces from browser automation sessions into NDJSON.

Community
Advanced
SignmanalSignmanal
1

p5js

Create p5.js browser visuals with export to PNG, GIF, MP4, and SVG.

Community
Advanced