The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The scripts fetch_page.py, extract_links.py, and extract_text.py use the requests library to fetch content from external URLs provided as arguments at runtime. This is standard functionality for a web scraping tool.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external websites, which creates a surface for indirect prompt injection if the agent interprets the output as instructions.
Ingestion points: Data is ingested via requests.get() in all three scripts when fetching user-supplied URLs.
Boundary markers: The scripts output the scraped content to stdout or a file without explicit boundary markers or warnings to the agent that the content is untrusted.
Capability inventory: The skill has the capability to perform network requests (requests.get) and write content to the local file system (via the --output flag in extract_text.py and fetch_page.py).
Sanitization: The script extract_text.py proactively removes potentially problematic HTML tags, including <script>, <style>, <nav>, <footer>, <header>, and <aside>, before providing the text to the agent.

web-scraper