meeting-secretary

专业的会议秘书技能，用于分析会议转录文本并生成结构化会议纪要。当用户提供会议转录文本（或会议笔记文本），并要求分析会议内容、生成会议纪要、提取关键信息时使用此技能。

This skill appears coherent and implements what it claims: analyze meeting transcripts and produce structured minutes. Things to consider before installing/use: - The included script (scripts/split_long_transcript.py) is a simple, readable helper that splits long transcripts and writes segment files to disk—expect files to be created in the specified output directory. Review and run it in a safe workspace (not as root). - Documentation references additional scripts (transcribe_audio.py, process_video.py) and integrations (Notion/飞书) that are NOT included. If you need audio/video processing or automatic uploads, you'll need to obtain or implement those scripts and supply any API keys yourself. - The README suggests using packages like whisper or moviepy for optional features; those are not installed automatically. Installing those packages and running transcription will have their own requirements and privacy implications. - The skill processes potentially sensitive meeting content. Do not feed confidential transcripts unless you are comfortable with how outputs and any local files are stored and with any downstream services you integrate. Recommendation: you can install and use this skill for text-only transcript analysis after reviewing the script. If you plan to use audio/video processing or service integrations, verify those additional tools and required credentials separately and inspect any third-party code before running.

Type: OpenClaw Skill Name: meeting-secretary Version: 1.0.0 The 'meeting-secretary' skill bundle is a legitimate tool designed for analyzing meeting transcripts, identifying participant sentiments, and generating structured minutes. The provided Python script, `scripts/split_long_transcript.py`, is a utility for segmenting large text files based on timestamps or speakers using standard libraries and safe file handling practices. While the documentation in `references/best_practices.md` refers to additional scripts (transcribe_audio.py, process_video.py) not included in the bundle, the existing code and the AI instructions in `SKILL.md` focus entirely on the stated analytical purpose without any evidence of data exfiltration, malicious execution, or unauthorized system access.