The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions suggest using 'sudo' to install system-level packages such as 'curl', 'jq', and 'libxml2-utils'. It also utilizes shell heredocs to execute Python scripts embedded within the skill for XML processing.
[EXTERNAL_DOWNLOADS]: The skill fetches torrent metadata and magnet links from an external third-party API located at 'https://bitmagnetfortheweebs.midnightignite.me/torznab/api', which is an untrusted source.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests and processes untrusted XML data from a remote API. * Ingestion points: XML response body from the Torznab API. * Boundary markers: There are no delimiters or 'ignore embedded instructions' warnings applied to the data before it is presented to the agent or user. * Capability inventory: The skill can execute network requests via 'curl' and run code via 'python3' and 'npm'. * Sanitization: No sanitization or validation is performed on the 'title' or other metadata fields returned by the external service before they are processed.

torrent-search