pinchtab

Automate local browser interactions via accessibility trees and stable element references.

Updated Aug 23, 2026
One-click install
npx skills add https://github.com/escotilha/claude-public --skill pinchtab-escotilha
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: pinchtab
Source: https://github.com/escotilha/claude-public/tree/main/skills/pinchtab
Command: npx skills add https://github.com/escotilha/claude-public --skill pinchtab-escotilha

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

PinchTab eliminates expensive, brittle browser automation by letting agents interact with pages using accessibility trees and stable element references instead of relying on heavy screenshot-based grounding.

Core Features & Use Cases

  • Browser automation via local HTTP + CLI: Drives a Chrome instance through a lightweight control server for repeatable workflows.
  • Accessibility-tree snapshots with stable element refs: Captures interactive elements and page text in token-efficient formats to enable reliable clicking, filling, and navigation.
  • Verification and exports: Supports screenshots, PDF export, and JavaScript evaluation for inspection and troubleshooting.
  • Profiles and multi-agent safety: Keeps cookies/sessions across runs, includes tab locking, and offers stealth modes for safer parallel operation.

Quick Start

Run pinchtab to navigate to a URL, snapshot interactive elements, click and fill referenced inputs, and snapshot again to confirm the results.

Frequently Asked Questions about pinchtab

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

FAQPage Schema
How do I automate browser tasks without using expensive screenshot grounding?

You can automate browser interactions using accessibility tree snapshots with stable element references instead of screenshot grounding. This token-efficient method captures interactive elements and page text for reliable clicking and filling.

What is an accessibility tree snapshot for browser automation?

An accessibility tree snapshot captures interactive elements and page text in a token-efficient format with stable element references. It enables reliable clicking, filling, and navigation without relying on heavy screenshot-based visual grounding.

Can I run authenticated browser sessions in parallel for QA testing?

Authenticated session automation supports parallel operation through persistent session profiles that retain cookies across runs. Tab locking and stealth modes ensure multi-agent safety during concurrent QA testing workflows.

How do I extract structured data from multi-page web applications?

You extract structured data from multi-page web applications by navigating via a local HTTP API, snapshotting interactive elements and text, and evaluating JavaScript. Stable element references ensure reliable targeting across multiple pages.

Does browser automation via local HTTP API support PDF and screenshot export?

Browser automation via local HTTP API supports screenshot and PDF export for inspection and troubleshooting. It also provides CLI actions for navigation, snapshotting, clicking, filling, pressing, and JavaScript evaluation.

When should I avoid screenshot-based browser automation?

You should avoid screenshot-based browser automation when tasks require token efficiency, repeatable workflows, or reliable element targeting across authenticated sessions. Accessibility tree snapshots provide a stable, lightweight alternative for form filling and QA testing.