agentation-self-driving

Overview

Agentation is an agent-agnostic visual feedback tool that bridges the gap between what users see on their web page and what AI coding agents can understand. Users can click elements, select text, drag-select regions, and add notes directly on their page. The tool then generates structured markdown output with CSS selectors, positions, and context that AI agents can use to find the exact code being referenced.

Key Features

• Click-to-annotate with automatic selector identification: Click any element on the page and Agentation automatically identifies its CSS selector for precise code location.
• Multiple selection modes: Supports individual element clicks, text selection for specific content, multi-select via drag for multiple elements, and area selection for annotating any region including empty space.
• Animation pause capability: Freezes all CSS animations, JavaScript animations, and videos to capture specific visual states for feedback.
• Structured output for AI agents: Generates markdown with selectors, positions, and context that AI agents can grep for in the codebase.
• Zero runtime dependencies: Built with pure CSS animations and no external libraries, keeping the installation lightweight and conflict-free.

Who is this for?

This skill is designed for developers who use AI coding agents and want a faster way to point at specific UI elements when requesting changes. Instead of describing elements in text, you can visually annotate your page and copy structured output that any AI agent can understand and act on.