Wjs Transcribing Audio
PassAudited by ClawScan on May 13, 2026.
Overview
The skill is coherent for making SRT transcripts, but users should notice that it may send media to cloud ASR providers and use API credentials.
This looks appropriate for a transcription skill. Before installing, make sure you are comfortable uploading the selected audio/video to OpenAI or Volcano, using an API key that may incur costs, and letting the skill choose providers by language unless you specify otherwise.
Findings (4)
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.
Your OpenAI account key may be used to upload audio for transcription and may be billed by the provider.
The skill uses a user's OpenAI API key to call the transcription API. This is purpose-aligned, but it is sensitive account authority and can incur usage charges.
headers={"Authorization": f"Bearer {os.environ['OPENAI_API_KEY']}"}Use a dedicated or restricted API key where possible, monitor usage, and revoke the key if you no longer use the skill.
Audio or video content selected for transcription may be sent to OpenAI or Volcano depending on the source language.
The skill explicitly routes audio transcription through external ASR providers. That is central to the stated purpose, but it means user media content may leave the local environment.
Chinese defaults to Volcano (豆包) ASR; other languages ... use OpenAI Whisper API
Do not use this skill with media you are not allowed to share with the selected ASR provider; confirm the provider choice for sensitive recordings.
The agent may process local media files and create temporary audio chunks as part of transcription.
The skill documents local media conversion with ffmpeg. This is expected for transcription preparation, but it reads local media files and writes chunk files.
ffmpeg -hide_banner -loglevel error -y \ -ss <start> -t 600 -i input.mp4 \ -vn -ac 1 -ar 16000 -c:a libmp3lame -b:a 64k chunk.mp3
Run it only on intended input files and keep output paths in a workspace you control.
Chinese audio may be routed to Volcano by default based on a stored preference, even if you would prefer another provider.
The skill references persistent user memory or a standing preference to choose the ASR provider for Chinese audio. This is not inherently unsafe, but stale or incorrect memory could affect where media is sent.
This is a deliberate routing decision — see user's memory on Chinese ASR priority.
Explicitly state the ASR provider you want when handling sensitive or provider-restricted media.