The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Python subprocess module to execute system binaries for media processing tasks.
Evidence: Found in scripts/media_convert.py, scripts/batch_resize.py, and scripts/video_optimize.py where the scripts invoke ffmpeg, ffprobe, and magick.
Note: Commands are executed using argument lists, which is a recommended security practice to prevent shell injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill processes external media files, which represents a potential surface for indirect prompt injection through malicious file metadata or specially crafted filenames.
Ingestion points: scripts/batch_resize.py, scripts/media_convert.py, and scripts/video_optimize.py accept file paths as input.
Boundary markers: The skill does not implement explicit boundary markers for untrusted data.
Capability inventory: The skill has the capability to write to the local file system and execute subprocesses.
Sanitization: The implementation uses pathlib.Path objects and list-based subprocess execution, which significantly reduces the risk of traditional command injection.

media-processing