The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the subprocess module to execute ffmpeg and ffprobe for audio duration detection, segment extraction, and final clipping. All command invocations use list-based arguments rather than shell strings, which effectively prevents command injection vulnerabilities.
[EXTERNAL_DOWNLOADS]: The transcription logic relies on the FunASR library, which fetches pre-trained models (e.g., paraformer-zh) from the Alibaba ModelScope repository. This is standard behavior for machine learning tools and utilizes a well-known, established service.
[COMMAND_EXECUTION]: The main script supports an optional integration with a separate voice-changer skill. It executes the external script using subprocess.run with list arguments, ensuring that user-supplied parameters for the voice type are not interpreted as shell commands.

audiocut-keyword