The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructions include installing the openai-whisper package from the Python Package Index and torch from the official PyTorch registry.
[COMMAND_EXECUTION]: The skill uses the whisper command-line tool to perform local audio transcription.
[PROMPT_INJECTION]: The skill creates an indirect prompt injection surface by converting untrusted audio input into text that is then processed by the AI agent without sanitization. Ingestion points: Audio files are ingested and transcribed as shown in the workflow section of SKILL.md. Boundary markers: The skill does not implement delimiters or instructions to ignore embedded commands within the transcribed text. Capability inventory: The agent has the capability to execute shell commands via the whisper CLI and run Python code. Sanitization: No sanitization or validation of the transcribed text is performed before it is used in subsequent agent steps.

whisper-stt