The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches data from external RSS and Atom feeds using curl. While several examples target well-known and trusted sources like GitHub and Hacker News, the skill is functionally designed to fetch content from any URL provided by the user or found in data.
[COMMAND_EXECUTION]: The skill employs bash -c for executing shell commands and explicitly instructs the agent to concatenate variables into command strings (e.g., bash -c 'curl -s "'"$feed"'"'). This pattern constitutes a command injection vulnerability if the variables contain unsanitized user-provided input.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection (Category 8) because it ingests untrusted XML content from the web and processes it without using boundary markers or content sanitization.
Ingestion points: Remote RSS/Atom feeds retrieved via curl (SKILL.md).
Boundary markers: None specified in the instructions to delimit external content.
Capability inventory: The skill has access to powerful shell utilities including curl, xmllint, awk, and bash.
Sanitization: No sanitization or validation of the fetched XML content is performed before the agent processes it.

rss-fetch