The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references standard machine learning libraries such as openai-whisper, transformers, and torch. These packages are maintained by trusted organizations.
[COMMAND_EXECUTION]: Instructions are provided for installing ffmpeg via system package managers and for using the whisper CLI. These are standard operations for the tool's functionality.
[INDIRECT_PROMPT_INJECTION]: The skill processes audio data into text, representing a data ingestion surface.
Ingestion points: Audio files are ingested via model.transcribe() in SKILL.md and references/languages.md.
Boundary markers: Absent; there are no instructions to disregard potential commands within the audio.
Capability inventory: Examples include writing output to the file system using open().write() in SKILL.md.
Sanitization: Absent; the skill does not perform validation or filtering on the transcribed content.

whisper