The Agent Skills Directory

[SAFE]: The skill instructions and provided code snippets perform legitimate media analysis tasks without any malicious intent or hidden behaviors. No persistence mechanisms, obfuscation, or hardcoded credentials were found.
[EXTERNAL_DOWNLOADS]: The code utilizes the mediabunny library's UrlSource to fetch media files from remote URLs for processing. The examples provided use the project's own domain (remotion.media), which is a standard and expected practice for documentation.
[INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface for indirect prompt injection as it processes untrusted media files from user-supplied URLs. However, the implementation is standard for media analysis tools and does not provide a path for the agent to execute embedded instructions. Sanitization is handled by the underlying media library.

remotion-multimedia