CyberKurry

Video content analysis pipeline — extract frames, transcribe audio, run LLM visual+audio analysis, synthesize structured Bilibili publish metadata (title, intro, tags, category, cover suggestion). Use when user says '生成投稿元数据', '视频元数据分析', or explicitly requests this skill. ⚠️ API modes send video frames and audio to external LLM providers — see Privacy section.

Generate or edit images with Volcengine Seedream (Doubao). Use for image creation requests incl. text-to-image, image-to-image, multi-reference fusion, sequential/group generation, PNG output, prompt optimization, web search, streaming. Supports Seedream 5.0-lite / 4.5 / 4.0.