The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill fetches content from external RSS/Atom feeds, which serves as a potential surface for indirect prompt injection if the feed content is malicious.
Ingestion points: Data is ingested from various external URLs via curl in SKILL.md, including Hacker News, Reddit, and TechCrunch.
Boundary markers: There are no explicit boundary markers or instructions to the agent to isolate the fetched content from its own instructions.
Capability inventory: The skill utilizes shell utilities such as xmllint, grep, awk, and tr for parsing and formatting.
Sanitization: The skill does not perform sanitization of the fetched text (titles, descriptions, links) before processing it.
[COMMAND_EXECUTION]: The skill uses shell commands to interact with the web and process data.
Uses curl to fetch XML and Atom data from remote URLs.
Uses xmllint to parse and query XML data via XPath.
Uses standard text processing tools like awk, grep, and tr to manipulate the feed output.
[EXTERNAL_DOWNLOADS]: The skill fetches content from multiple external domains to retrieve news feeds.
Retrieves data from well-known services and platforms such as Hacker News (hnrss.org), TechCrunch (techcrunch.com), Ars Technica (arstechnica.com), and Reddit (reddit.com).
Fetches blog updates from GitHub's official feeds.

rss-fetch