WeChat Voice

专为微信 clawbot 设计的微信语音解析技能。识别微信 SILK 语音、解码为 WAV、转写为文字，并基于语音内容回复。

What this skill does

Inspect an inbound audio attachment.
Detect whether it is WeChat SILK (#!SILK_V3).
Decode SILK audio to WAV with Python.
Transcribe WAV locally with CPU-based Whisper.
Return a transcript or explicitly report that the clip is blank / too short / unclear.
Use the transcript as the basis for the user-facing reply.