The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (LOW): The skill fetches content from external URLs via the markdown.new proxy. While the proxy is not on the trusted sources list, the skill includes explicit safety rules to mitigate SSRF by blocking local, internal, and metadata IP addresses.
[PROMPT_INJECTION] (LOW): The skill possesses a surface for indirect prompt injection (Category 8) as it retrieves and processes arbitrary content from the internet.
Ingestion points: Web content retrieved via WebFetch from user-provided URLs.
Boundary markers: None explicitly defined to separate fetched content from system instructions during processing.
Capability inventory: Uses WebFetch to access external data.
Sanitization: Implements URL-level sanitization (protocol checks, IP blacklisting, token detection) but lacks content-level sanitization for the fetched markdown.
[SAFE] (SAFE): No hardcoded credentials, obfuscation, or persistence mechanisms were detected. The skill is transparent about its operational flow.

markdown-fetcher