The Agent Skills Directory

[SAFE]: No malicious patterns, obfuscation, or security vulnerabilities were detected. The skill consists primarily of documentation and best practices for the Remotion video framework, authored by the official vendor.
[INDIRECT_PROMPT_INJECTION]: The skill documents features that ingest external data (such as fetching JSON for composition metadata or Lottie animations). While these represent potential attack surfaces for indirect prompt injection if the sources are untrusted, they are inherent and documented features of the platform.
Ingestion points: calculate-metadata.md (demonstrates fetching from props.dataUrl), lottie.md (demonstrates fetching Lottie JSON data).
Boundary markers: None specified in the documentation examples.
Capability inventory: The skill references capabilities including file system access via scripts, network operations, and execution of external tools like FFmpeg.
Sanitization: The provided examples do not demonstrate explicit sanitization of the fetched JSON data, which is expected for instructional material.
[EXTERNAL_DOWNLOADS]: The skill references downloading assets and scripts from vendor-owned domains and well-known, trusted service providers.
Trusted Sources: References include Google Fonts, MapTiler, Mapbox, ElevenLabs, and MapLibre (via unpkg CDN).
Context: These downloads are neutrally documented as part of the setup for fonts, map tiles, and AI-generated voiceovers, representing standard development workflows.

remotion-markup