voice-to-text
Pass
Audited by Gen Agent Trust Hub on Jul 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a standard voice-to-text workflow using legitimate, well-known libraries such as Vosk and the ffmpeg utility.
- [EXTERNAL_DOWNLOADS]: Instructions provided for downloading language models target alphacephei.com, which is the official website and repository for the Vosk project. This is a well-known service for this specific technology.
- [COMMAND_EXECUTION]: The script executes ffmpeg to convert audio files into a compatible format. These calls use the subprocess module with a list of arguments rather than a raw shell string, effectively mitigating shell injection risks even when processing user-supplied file paths.
- [DATA_EXFILTRATION]: No network operations beyond the initial model downloads were detected. The speech recognition process itself is performed entirely offline on the local machine.
Audit Metadata