If you're building automation that needs to actually join Zoom meetings, capture media streams, or react to live session data, this is your starting point. It walks you through bot architecture decisions like whether you're observing, transcribing, or taking action, then routes you to Meeting SDK and RTMS for the core implementation. The skill's main value is helping you avoid the classic mistakes: confusing batch transcription with live media workflows, skipping the auth model design, and forgetting about post-meeting storage. It's less a tutorial and more a decision tree that keeps you from painting yourself into corners with Zoom's bot lifecycle constraints.
npx skills add https://github.com/anthropics/knowledge-work-plugins --skill build-zoom-bot