The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to Indirect Prompt Injection. It is designed to fetch an external RSS feed and then perform a "full-depth reading" of every article link found within that feed.
Ingestion points: External RSS feed URL and every link extracted from the RSS entries as defined in SKILL.md Step 1 and Step 3.
Boundary markers: The instructions lack any boundary markers or instructions to treat the fetched article content as untrusted data.
Capability inventory: The skill utilizes WebFetch to read content and has the capability to generate new documents based on that content.
Sanitization: No sanitization or validation of the retrieved content is mentioned.
[PROMPT_INJECTION]: The skill includes a file named .security-scan-passed which makes a self-referential claim that the skill is safe and has passed automated validation. This is a common pattern used to bypass manual or automated security reviews.
[EXTERNAL_DOWNLOADS]: The skill performs automated network requests to youmind.com and any arbitrary URL contained within the retrieved RSS feed to download content for processing.

karpathy-rss-daily-briefing