PPT image-design workflow from uploaded source materials, text, and optional visual references to reusable visual preview directions, confirmed high-resolution slide images, and image-based PPTX output. Use when Codex needs to create or redesign a presentation, generate PPT outlines, cover/homepage content, visual collage方案, Image2 visual exploration, style reuse, template沉淀, or 图片版PPT.
Default replacement image generation and image editing skill for Codex API-key environments where the built-in imagegen tool is unavailable. Use for any request to create images, generate visual assets, make illustrations, render concepts, create product/mockup images, edit or restyle images, or generate from reference images, unless the user explicitly requires another provider or a repo-native vector/code asset.
Generate long-form infographic images with HTML/CSS and Playwright, optimized for whitespace, reading rhythm, and low-density information design instead of short-card carousel layouts.
帮助用户创作并生成小红书信息密度高、媒体感强的数据图文与图文轮播图片。适用于用户提到小红书图文、科技图文、数据图文、图文卡片、轮播图、封面页、信息图、趋势图解、观点型图文、对比型图文、行业拆解图文、知识博主风图文等场景。先锁定主题、受众、观点和证据，再根据内容类型自动选择合适的主题模板与页数组织方式，生成可执行导演稿，然后产出 HTML/CSS 页面并渲染成最终图片，而不是只给零散文案。
Use when the user wants to convert Markdown or structured written content into a polished PDF with good typography, cover metadata, optional table of contents, and print-ready layout. Trigger on requests like "转成 PDF", "排版成 PDF", "导出 PDF", "做成报告 PDF", "把内容排好版", or when the main need is formatting and exporting content rather than writing the content itself.
Use when the user wants to write a Chinese technical book, handbook, long-form guide, playbook, or "from beginner to mastery" manual with strong narrative structure, chapter rhythm, comparisons, scenarios, and actionable conclusions. Trigger on requests like "写一本技术书", "从入门到精通", "技术手册", "方法论书稿", "出一本 PDF 书", or when asked to turn a technical topic into a structured manuscript.