openai-whisper
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No security issues detected. The skill facilitates the use of the official OpenAI Whisper tool for local speech-to-text processing.
- [EXTERNAL_DOWNLOADS]: Mentions that the whisper CLI downloads pre-trained models from OpenAI's official infrastructure to ~/.cache/whisper upon first execution. This is the documented and expected behavior for this tool.
- [COMMAND_EXECUTION]: Instructs the agent on how to call the local whisper binary to perform audio transcription and translation, which is the primary intended function of the skill.
Audit Metadata