@liyayong000-sketch

zeelin-text-analysis performs structural extraction on uploaded text files via a three-step api workflow (upload, submit, poll), with mandatory status polling and markdown result formatting. requires zeelin platform credentials and handles documents up to 500mb.

zeelin-video-analysis extracts entities and generates relationship graphs from uploaded video files through a three-step api workflow (upload, submit, poll), but lacks clarity on failure handling and result schema.

剧本生成技能V4，将文本文件（小说/故事）转换为完整的影视分镜剧本。当用户说'生成剧本'、'小说转剧本'、'创建分镜'、'影视改编'等并上传文本文件时使用此技能。