The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and display untrusted content from external web sources, which is a primary vector for indirect prompt injection attacks.
Ingestion points: In scripts/feed.py, external feed data is retrieved from user-supplied URLs via the urllib.request.urlopen method.
Boundary markers: The skill body and instructions do not provide the agent with specific boundary markers or instructions to treat the retrieved feed content as data rather than instructions.
Capability inventory: The skill's script is limited to network retrieval (fetching feeds) and text processing. It lacks dangerous capabilities such as file system writing, subprocess execution, or dynamic code evaluation.
Sanitization: The script includes a strip_html function that removes HTML tags from entry titles and summaries using regular expressions, and it truncates entry summaries to a maximum of 2000 characters.

rss-feeds