meeting record analysis

将会议录音转成结构化会议纪要。适用于用户上传会议音频后，希望通过 ASR 转写、LLM 总结和可选 TTS 播报，自动提取会议主题、讨论要点、决策和行动项的场景。输入支持 `audio_file`、`need_voice_summary`、`language`；默认输出 JSON 结构化纪要，并可附带语音摘要文件路径。

Before installing or running this skill: (1) Recognize that audio and transcript text are uploaded to external endpoints (models.audiozen.cn and api.senseaudio.cn by default). Only use with recordings you are comfortable sending off‑host. (2) The script expects API keys (LLM, ASR, TTS) via environment variables or .env files, but the registry metadata did not list these — confirm which secrets you'll provide and trust the service operators. (3) Inspect the .env.example and the script to confirm which env names and endpoints will be used; consider overriding endpoints to services you trust. (4) Run in an isolated environment (container) and review outputs/ for any sensitive intermediate files; remove credentials from working dirs when finished. (5) If you need higher assurance, ask the publisher for verified source/homepage and for the skill metadata to declare required env vars and endpoints explicitly.

Type: OpenClaw Skill Name: meeting-record-analysis Version: 1.0.0 The skill is a legitimate meeting minutes assistant designed to automate audio transcription, text cleaning, and summarization using external AI services (SenseAudio and AudioZen). The core logic in `scripts/run_meeting_minutes.py` follows a standard workflow: it transcribes audio via an ASR API, cleans the text using an LLM, and generates a structured summary in JSON format. The script uses well-known libraries like `openai` and `requests` and handles API keys through environment variables. No evidence of malicious intent, data exfiltration, or dangerous vulnerabilities was found; all network and file operations are directly aligned with the stated purpose of processing meeting recordings.