skills.rest
ISAQQSAI
ISAQQSAICOMMUNITY
Explore all skills by this author →
GitHub

agent-browser

Automate headless browser interactions using ref-based selectors and accessibility-tree snapshots.

Updated Mar 11, 2026
Search Tags:#automation#browser#headless#refs#ai-agents#web-automation#accessibility-tree
One-click install
npx skills add https://github.com/ISAQQSAI/SkillAttack --skill agent-browser-isaqqsai
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/ISAQQSAI/SkillAttack/tree/main/data/hot100skills/009_matrixy_agent-browser-clawdbot
Command: npx skills add https://github.com/ISAQQSAI/SkillAttack --skill agent-browser-isaqqsai

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Headless browser automation is essential for AI agents to perform web interactions reliably without a visible UI, enabling reproducible actions and deterministic element targeting.

Core Features & Use Cases

  • Deterministic element interaction using accessibility-tree snapshots and ref-based selectors for stable automation in complex SPAs.
  • Session isolation and multi-tab support to run parallel experiments or agent simulations without bleed-over.
  • Snapshotting and state management to capture page state for debugging, auditing, and reproducibility across runs.

Quick Start

Run a quick smoke test to verify the agent-browser workflow with a sample page.

Frequently Asked Questions about agent-browser

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

FAQPage Schema
How do I automate headless browser interactions deterministically for AI agents?▼

Headless browser automation for AI agents is handled deterministically by using accessibility-tree snapshots and ref-based selectors to target UI elements reliably. This enables stable automation in complex SPAs and supports multi-step interactions without a visible UI.

What is ref-based element targeting in web automation?▼

Ref-based element targeting in web automation uses references generated from accessibility-tree snapshots to interact with UI elements. This approach ensures deterministic, stable targeting across dynamic web apps, avoiding the fragility of traditional DOM selectors.

Can I run parallel web automation experiments with isolated sessions?▼

Parallel web automation experiments are supported through session isolation and multi-tab management. This allows you to run concurrent agent simulations or workflows without state bleed-over, ensuring reproducible results across isolated runs.

How do I capture page state for debugging during browser automation?▼

Capturing page state during browser automation is achieved through snapshotting and state management features. These tools capture the page state at specific steps, enabling debugging, auditing, and reproducibility across multiple automation runs.

Does this browser automation approach work with single page application workflows?▼

Single page application workflows are fully supported by using accessibility-tree snapshots and ref-based selectors. This combination enables reliable, multi-step interactions across dynamic SPA content, ensuring deterministic automation results.

Related Skills

Explore More Software Engineering Skills→
daichan132daichan132
34

agent-browser

Automate browser interactions and extract data via CLI.

Community
Advanced
MudassarAbrarMudassarAbrar

agentic-eval

Evaluate and refine AI agent outputs through iterative generate-evaluate-critique-refine loops.

Community
Advanced
zhouxiuminzhouxiumin

agent-browser

Automate browser workflows using accessibility-tree snapshots and ref-based element targeting.

Community
Advanced
Mte90Mte90
51

core-actionbook

Retrieve structured browser automation action manuals by ID and keyword.

Community
Intermediate