The Agent Skills Directory

[SAFE]: The skill facilitates the use of the OpenAI Whisper command-line tool. All references are to official sources (openai.com).
[EXTERNAL_DOWNLOADS]: The skill notes that Whisper models are downloaded upon first execution to ~/.cache/whisper. This is standard behavior for this well-known tool.
[COMMAND_EXECUTION]: The skill provides examples of running the whisper CLI with various parameters for audio processing. This is consistent with its stated purpose of providing a local speech-to-text capability.
[INDIRECT_PROMPT_INJECTION]: The skill defines an interface for processing external audio files via file paths. It documents the ingestion points for these files and the resulting capability to generate text transcriptions.

openai-whisper