This handles rich media in QQBot conversations by letting Claude send and receive images, audio, video, and files through a simple `<qqmedia>` tag wrapper. It auto-detects media type from file extensions, works with both local paths and URLs, and automatically downloads incoming images so you can reference them directly. The file size checks are baked in (30MB for images, 100MB for video/files), which saves you from hitting API limits. It's built for the openclaw framework and needs QQBot channels configured. The speech-to-text integration for voice messages is a nice touch if you have STT enabled, though it defaults to just passing through the audio path otherwise.
npx skills add https://github.com/sliverp/qqbot --skill qqbot-media