The Agent Skills Directory

[SAFE]: The skill implements media processing logic through a collection of Python and Shell scripts. Technical analysis confirms that all external utility calls to ffmpeg and ffprobe are performed using secure argument lists, precluding command injection vulnerabilities.
[COMMAND_EXECUTION]: The skill automates complex command-line operations for media analysis and encoding. The usage of subprocess.run across scripts like probe-media.py, detect-segments.py, and smart-compress.py is verified to be safe and limited to the intended media processing functions.
[INDIRECT_PROMPT_INJECTION]: While the skill ingests external data from media metadata and JSON configuration files (EDL), it does so within a strictly technical scope. There is no evidence of an exploitable surface that would allow untrusted data to manipulate the agent's high-level instructions.

ffmpeg-ops