Voice To Text

• Verdict: Pass
• Confidence: High
• Analyzed: Feb 17, 2026, 6:36 AM
• Findings: 0
• Version: 1.0.0

Overview

The skill's requirements, code, and instructions are consistent with an offline Vosk-based transcription tool; nothing requested is disproportionate to its stated purpose.

This skill appears to do exactly what it says: offline transcription using Vosk and ffmpeg. Before installing: (1) be prepared to install ffmpeg (brew/apt) and the Python vosk package (pip); note the top-level registry install entry only lists brew for ffmpeg, so you may need to pip install vosk yourself. (2) You will need to download Vosk model archives from alphacephei.com (these are large for higher-accuracy models) — verify you trust that source and have disk space. (3) The skill reads the audio files you give it and stores/reads models in ~/.vosk/models; it creates a temp WAV in /tmp while processing. (4) Once a model is downloaded, processing is offline. If you require stricter controls, review the model download URLs and install packages manually rather than relying on automated install hooks.