The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill requires the user to fetch and execute a collection of runtime scripts (JavaScript, Python, Swift, Bash) from a third-party GitHub repository (github.com/jack-reson8/pro-motion) to function. While it recommends using a pinned commit and obtaining user consent, executing unbundled code from an external source is a notable risk.
[COMMAND_EXECUTION]: Extensive use of shell commands to process media, including node for parallel rendering, python3 for audio analysis, and bash for reference frame extraction.
[EXTERNAL_DOWNLOADS]: Fetches external resources including code, fonts, and video references from various web sources such as YouTube and Behance.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data, such as video URLs and audio files, which could contain instructions designed to influence the agent's behavior.
Ingestion points: External video URLs and audio files processed by ref_frames.sh and analyze_music.py.
Boundary markers: None identified in the provided instructions to delimit external content from internal logic.
Capability inventory: Includes subprocess calls to node, python3, bash, swift, and ffmpeg across multiple scripts.
Sanitization: No explicit sanitization or validation of external URLs or media metadata is documented.

pro-motion