agent-browser

Automate browser navigation and interactions using accessibility tree snapshots and ref-based element addressing.

Updated Aug 27, 2026
One-click install
npx skills add https://github.com/eninem123/hunterclaw --skill agent-browser-eninem123
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: agent-browser
Source: https://github.com/eninem123/hunterclaw/tree/main/skills/agent-browser-clawdbot
Command: npx skills add https://github.com/eninem123/hunterclaw --skill agent-browser-eninem123

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

You need reliable, repeatable browser interactions for multi-step web workflows, but visual or heuristic element selection often breaks when the page changes.

Core Features & Use Cases

  • Accessibility tree snapshots: Capture a structured view of the page for AI-friendly decision making.
  • Ref-based deterministic element selection: Interact with stable element references to reduce flaky automation.
  • Session isolation & state persistence: Run separate browser contexts and reuse saved auth state to skip logins.
  • Use Case: Automate a complex SPA workflow like navigating to a dashboard, selecting options, submitting a form, and extracting results using refs and re-snapshots after each UI change.

Quick Start

Open your target URL, snapshot interactive refs with agent-browser snapshot -i --json, then click and fill elements using the returned refs in a second step.

Frequently Asked Questions about agent-browser

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

FAQPage Schema
How do I automate browser interactions without flaky element selectors?

Browser automation can be made deterministic by capturing accessibility tree snapshots and interacting with elements using stable ref-based addressing. This approach avoids visual or heuristic selection, ensuring repeatable multi-step workflows even when the page UI changes.

What is accessibility tree snapshot-based web testing?

Accessibility tree snapshot-based web testing captures a structured, AI-friendly view of a webpage's interactive elements. It generates JSON output with unique refs, allowing automation scripts to target elements deterministically rather than relying on fragile CSS selectors.

Can I run isolated browser sessions for automated form filling?

Yes, you can run isolated browser sessions for automated form filling. The tool supports session isolation and state persistence, allowing you to maintain separate browser contexts and reuse saved authentication states to skip logins during testing.

How do I extract text and attributes from a complex SPA?

To extract text or attributes from a complex SPA, you snapshot the page using the headless CLI with JSON output, perform interactions like clicking or filling by referencing the returned @ids, and then re-snapshot after each UI change to extract updated results.

Does deterministic browser automation work with headless CLI workflows?

Yes, deterministic browser automation works with headless CLI workflows. You execute commands with flags like -i and --json to capture interactive refs, enabling stable element targeting and multi-step navigation without launching a visible browser interface.