---

name: crawl4ai-skill description: Integrates with web crawling libraries to ingest live documentation.

Crawl4AI Skill

This skill allows the agent to ingest and summarize documentation from external websites using its read_url_content capabilities, effectively simulating a "crawling" behavior for knowledge acquisition.

Usage

When the user gives a URL and asks to "learn from this" or "read the docs":

1. Fetch Content: Use read_url_content to get the markdown representation of the page.
2. Summarize/Ingest:
   • If the content is a single page, analyze it and summarize key API points, configuration options, or concepts relevant to the user's task.
   • If the content is a landing page, identify the most relevant sub-links (e.g., "Getting Started", "API Reference") and offer to read those next.
3. Documentation Generation: optionally save the summarized knowledge into a local markdown file (e.g., docs/external/lib_name.md) for persistent reference.

Prompting Strategy

• "Reading [URL] to extract API signatures..."
• "Ingesting documentation for [Library Name]..."