Audio Transcribe
PassAudited by ClawScan on May 1, 2026.
Overview
This skill appears to do what it claims: locally transcribe audio files, with the main things to notice being its external Python/model downloads and access to voice-message contents.
This appears suitable if you want local voice-message transcription. Before installing, make sure you are comfortable installing faster-whisper and downloading Whisper models locally, and remember that any voice message processed by the skill may be exposed as text in the agent conversation.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Installing the skill requires trusting the faster-whisper package and downloaded model files.
The skill relies on an external Python package and automatic model downloads, which are disclosed and central to local transcription but introduce normal dependency/provenance considerations.
pip install faster-whisper Models download automatically on first use.
Install dependencies from trusted package indexes, consider pinning versions, and review local storage/RAM needs before use.
Private audio content may become readable text in the agent interaction when the skill is enabled.
Incoming voice-message content is converted into text for the agent/tool output. This is the intended function, but the transcript may contain sensitive information or user-spoken instructions.
Clawdbot auto-transcribes incoming voice messages when this skill is enabled.
Enable this only where automatic transcription is desired, and treat generated transcripts as potentially sensitive user content.