The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill documentation includes instructions to install several standard Python libraries via pip for its core functionality:
httpx (HTTP client)
beautifulsoup4 and lxml (HTML parsing)
trafilatura (Article extraction)
markdownify (Markdown conversion)
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted content from the public web, which inherently presents an indirect prompt injection surface.
Ingestion points: Website content fetched via httpx in SiteCrawler.crawl and SiteCrawler.crawl_docs methods.
Boundary markers: The code does not demonstrate explicit boundary markers or instructions to ignore embedded commands within the extracted markdown/text before it is passed to the ingest function.
Capability inventory: The skill possesses network communication capabilities (httpx) and the ability to parse complex structured data (BeautifulSoup, ElementTree).
Sanitization: The skill employs trafilatura and markdownify for extraction, which naturally sanitizes content by removing scripts, styles, and headers/footers, significantly reducing the attack surface.

site-crawler