Agent Action Optics

Audits how ready a page is for AI agents: actionable elements, accessible names, operability, and an exportable readiness report.

Agent Action Auditor is a developer tool designed to enhance the visibility of interactive elements on any web page, making it easier to audit, debug, and design for agent-based automation systems like Atlas or other LLM-powered UI agents. This extension programmatically scans the current page to identify actionable DOM elements—such as buttons, inputs, links, and ARIA-labeled regions—and provides: Key Features: ★ Real-Time Scanning & Categorization Automatically detects and organizes actionable elements by type: buttons, text inputs, selects, links, textareas, and ARIA-accessible regions. ★ Highlight Overlay System Applies prominent, customizable highlight markers to each matched element with animated pulsing effects—ideal for visual debugging and accessibility audits. ★ Interactive Preview Panel Displays a collapsible list of detected elements within a sticky, scrollable popup. Clicking any entry scrolls and focuses the corresponding element in the DOM. ★ Dark/Light Theme Toggle ADA-compliant popup interface with a sticky header, sticky footer, and a built-in theme switcher to match your dev environment. ★ Manual Control & Filtering Enable or disable highlighting for specific categories of elements to streamline your testing and analysis workflows. ★ Context-Aware Design Built with modern web accessibility and agent compatibility in mind. Ideal for developers building browser-integrated agents, AI copilot experiences, or RPA tooling. Use Cases: ✅ Audit web UIs for agent compatibility ✅ Visualize what an AI or automation system "sees" ✅ Test interaction readiness for LLM-based browser assistants ✅ Assist in accessibility reviews for ARIA-rich interfaces ✅ Debug interface consistency across dynamic SPAs