@zmy1006-sudo

generates dark-themed architecture diagrams as self-contained html files with inline svg and semantic color coding for system components. covers design system specifics but lacks actionable generation procedure and error handling.

fmt-highdef-dynamic generates 4k static illustrations and 1080p educational videos for fmt microbiota transplant science content, with batch processing and style consistency features, but lacks clarity on actual execution paths and failure recovery.

fmt-video-production automates fecal microbiota transplant educational video creation by compositing video, subtitles, tts narration, and intros using ffmpeg. covers single and batch workflows with fallback strategies for missing dependencies.

baozong-trading frames stock analysis around buffett safety margins, munger checklists, and three-dimensional validation (fundamentals, technicals, capital flows). lacks executable procedures, data source integration, and failure mode handling.

xiaohongshu content production pipeline covering collection, curation, creation, image generation, and publishing. deployment instructions provided but workflow procedures lack numbered actionability and decision logic.

claude-obsidian-kb provides a three-layer vault architecture (raw/wiki/output) and triggers on pkm-related keywords, but lacks actionable step-by-step procedures, failure mode handling, and concrete implementation details.

通用流式AI检索问答技能 — 为任意行业应用提供四步流式分析交互界面。 触发场景：用户输入关键词 → AI自动执行：理解意图 → 检索知识库 → 流式生成 → 来源标记 → 完整回答。 当需要实现以下任意场景时激活： (1) AI搜索框 / 智能咨询组件重构 (2) 知识库问答（医疗/法律/金融/教育等垂直领域）...