traktor

What This Skill Does

Traktor is a high-performance web asset extraction tool for OpenClaw that functions as an intelligent, parallelized agent. Unlike basic scrapers, Traktor is designed for "paranoid-level" thoroughness. It automates browser navigation using MCP (Model Context Protocol) browser extensions to bypass common scraping hurdles like lazy-loading, dynamic JavaScript-rendered content, and complex DOM structures. When triggered, Traktor spawns sub-agents that execute controlled browsing sequences—including scrolling and viewport manipulation—to ensure every image, SVG, font, video, and stylesheet is captured and systematically organized into a structured local filesystem.

Installation

To add Traktor to your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/dorukardahan/traktor

Ensure that you have the claude-in-chrome MCP server configured and active, as the skill relies on these browser-specific hooks to interact with target websites.

Use Cases

Traktor is ideal for developers and designers who need to perform brand audits, competitive analysis, or site cloning for local development. Common scenarios include:

• Creative Asset Archiving: Quickly grabbing all source images, logos, and icons from a design portfolio.
• Front-end Auditing: Collecting CSS files and page structure to understand how a specific landing page layout is constructed.
• Asset Migration: Downloading all media from a legacy site to prepare for a migration project.
• UI/UX Research: Gathering site-wide component sets (icons, buttons, styles) for design system references.

Example Prompts

1. "/traktor https://stripe.com"
2. "Extract all assets from https://vercel.com and https://linear.app to my local folder."
3. "scrape website https://example.com for images and fonts."

Tips & Limitations

Traktor is extremely efficient but requires specific prerequisites.

Tips:

• Always ensure you have sufficient disk space when scraping multiple sites simultaneously, as the filesystem structure (pages, assets, styles) can grow quickly.
• Use the '/traktor' trigger with multiple URLs if you are performing a comparative analysis, as the agent is optimized to handle multiple parallel requests.

Limitations:

• Browser Dependency: This skill is strictly coupled to the mcp__claude-in-chrome__* tools. It will fail in headless environments that lack a compatible browser extension server.
• Anti-Bot Measures: While it mimics real user behavior (scrolling, loading), highly protected sites with advanced CAPTCHAs may still challenge the agent's navigation capabilities.