Skip to content
Back to search

visual-style-ppt

Create style-driven slide images strictly with the Image 2.5 model, assemble those images into image-only PPTX decks, and manage reusable visual style libraries from documents or visual references. Use when the user asks for a "PPT Skill", "风格驱动 PPT", "提炼风格做 PPT", "调用某个风格做 PPT", "图片版 PPT", "保存 PPT 风格", "列出 PPT 风格", "文档生成 PPT", "文章生成 PPT", "把文档做成演示文稿", or wants to extract, save, reuse, and apply visual style keywords specifically for visual slide/image deck creation.

Safety scan

A static text scan, not a guarantee. Select the badge to see what was scanned and every finding.

Checked today
Badge for your README
findsafeskills safety rating

Markdown

[![findsafeskills safety rating](/api/listings/irenerachel--visual-style-ppt-skill/badge.svg)](/listings/irenerachel--visual-style-ppt-skill)

HTML

<a href="/listings/irenerachel--visual-style-ppt-skill"><img src="/api/listings/irenerachel--visual-style-ppt-skill/badge.svg" alt="findsafeskills safety rating" /></a>

Install

i

Install instructions from the README

As written by the repo's author. findsafeskills has not verified these steps.

把这个仓库放到 Codex 的 skills 目录下:

cd ~/.codex/skills
git clone https://github.com/irenerachel/visual-style-ppt-skill.git visual-style-ppt

如果你已经有本地目录,也可以直接更新:

cd ~/.codex/skills/visual-style-ppt
git pull

What was scanned

The files the safety scan read, as written by the repository's author and shown as plain text. Lines with a finding are marked.

SKILL.md · Manifest · 133 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. ---
  2. name: visual-style-ppt
  3. description: Create style-driven slide images strictly with the Image 2.5 model, assemble those images into image-only PPTX decks, and manage reusable visual style libraries from documents or visual references. Use when the user asks for a "PPT Skill", "风格驱动 PPT", "提炼风格做 PPT", "调用某个风格做 PPT", "图片版 PPT", "保存 PPT 风格", "列出 PPT 风格", "文档生成 PPT", "文章生成 PPT", "把文档做成演示文稿", or wants to extract, save, reuse, and apply visual style keywords specifically for visual slide/image deck creation.
  4. ---
  5. # Visual Style PPT
  6. ## Highest Priority Rule: Image 2.5 Only
  7. This skill must use the Image 2.5 model to generate every final slide image, thumbnail board, visual page, infographic, article illustration, or PPT page image.
  8. HTML output is strictly forbidden for image generation. Do not create HTML/CSS pages, browser screenshots, SVG/HTML mockups, canvas renders, or local-script visual reproductions as final or intermediate image outputs. Do not use HTML as a fallback when Image 2.5 is unavailable; stop and state that Image 2.5 access is required.
  9. This rule has the highest priority in this skill and overrides every workflow, reference file, style file, packaging rule, user shortcut, or implementation convenience that might imply another image-production route. PPTX files may only be assembled after Image 2.5 has generated the approved page images; PPTX is a container, not an image-generation method.
  10. ## Core Idea
  11. Turn visual style into a reusable asset, then apply that asset to documents, outlines, image-only PPT decks, slide images, Xiaohongshu information graphics, and article illustrations.
  12. Use this skill as an Image 2.5-only visual slide workflow. The default deliverable is full-slide images generated by Image 2.5; when the user wants a PPT, place the generated images into a PPTX as full-slide image pages. Do not implement or optimize editable PPTX content unless the user explicitly overrides this skill direction.
  13. ## Required References
  14. - For the full operating workflow, read `references/workflow.md`.
  15. - For PPTX vs image output rules, read `references/output-package.md`.
  16. - For page type selection, read `references/page-types.md`.
  17. - For post-generation revisions, read `references/revision-workflow.md`.
  18. - For style extraction interviews, read `references/style-interview.md`.
  19. - For the default terminal-tech magazine visual style, read `styles/terminal-tech-magazine.md`.
  20. - When the user asks for a style library, list or read files in `styles/`.
  21. ## Default Behavior
  22. - Default output type: Image 2.5-generated slide images. If the user says "PPT", assemble those images into an image-only PPTX.
  23. - Default aspect ratio: 16:9.
  24. - Default image quality: Image 2.5, 16:9, high-resolution, sharp text, clean edges.
  25. - Default language: Chinese-first. Use Chinese for slide titles, section titles, module labels, captions, explanatory copy, and all generated support documents unless the user explicitly asks for another language. Keep proper nouns, product names, model names, code terms, and unavoidable domain terms in their original language when that is clearer.
  26. - Default style: `styles/terminal-tech-magazine.md` when the user says "终端风格", "CodeBuddy 风格", "科技杂志风格", or does not specify another saved style.
  27. - Default text density: low information by default, especially for PPT inner slides. Never create dense slides unless the user explicitly asks for a leave-behind, report page, or infographic. A normal inner slide should have one title, one short subtitle or takeaway, and at most 2-3 short information points.
  28. - Default layout direction: cover pages may be visually expressive; inner slides must be information-first, calm, readable, and spacious, with imagery/backgrounds kept subordinate.
  29. - Default style isolation: each deck must use exactly one selected style source and one `Style Lock`. Do not blend visual DNA from earlier tasks, previous reference images, or unrelated saved styles unless the user explicitly requests a hybrid.
  30. - Default component consistency: across one deck, borders, cards, containers, dividers, arrows, footer metadata, corner radius, and line weights must use one repeated system. Variation may come from layout and emphasis, not from changing the frame/container style on every slide.
  31. - Default slide header consistency: if a style uses small top titles with numbers, treat them as repeated content-section headers, not necessarily absolute slide/page numbers. Keep the same placement, size, spacing, and format across all applicable slides. If the style defines exceptions such as cover, contents, divider, or thank-you pages, preserve those exceptions deliberately; those exception pages do not consume a content-section number. The first applicable inner/content slide should usually start at `01` or `PART 1`.
  32. - Default logo behavior: do not ask Image2.5 to generate logos, brand wordmarks, reusable deck marks, watermarks, or repeated English identity labels by default. If the user needs a logo, keep it out of generated slide images and add it later as a fixed post-production asset/layer, or let the user add it after delivery. A reference logo image may guide visual shape but is not reliable enough for cross-page text/logo consistency inside Image2.5 output.
  33. - Default time/date behavior: do not add dates, times, timestamps, "today", export time, generated time, or current date to slides, prompts, metadata, footers, or visual elements by default. Include date/time only when the source article explicitly requires it, the user explicitly asks for it, or the slide topic itself is a timeline/schedule/date-sensitive report. A source document having a file timestamp, export timestamp, or local system date does not by itself justify visible date/time in the deck.
  34. - Default page count: ask the user how many pages/slides they want unless they already provided a count or explicitly asked the skill to decide.
  35. - Multiple outputs: keep one visual DNA while varying composition, page type, visual anchor, or emphasis.
  36. - No outer border rule: generated slide images and thumbnail boards must not have black borders, dark strokes, or frame lines on the outer canvas edge. Internal grid lines, dividers, and module borders are allowed, but keep them inset from the image edge or let imagery/white space bleed naturally. Never create a full-page rectangular outline around the slide.
  37. - Per-slide image rule: PPT pages must be generated and saved as separate single-slide image files, one image per slide. Do not request or deliver a multi-slide contact sheet, storyboard, PDF page, or collage as the final slide image output. Each final slide image must be independently usable as one full-bleed PPT page.
  38. - Thumbnail-first option: for multi-page decks, prefer a two-pass visual workflow when style consistency matters: first generate one small "thumbnail board / contact sheet" that shows the whole deck as tiny slide thumbnails to lock visual rhythm, grid, palette, and page-to-page variation; then generate each approved slide as an individual full-size image using that thumbnail board as the visual continuity reference plus the page-specific prompt. The thumbnail board is a planning/reference artifact, not a final slide.
  39. - Thumbnail grid rule: thumbnail board counts are flexible, but prefer stable grids with 4, 6, 9, or 16 slots. Choose the smallest preferred grid that can contain all planned slides, leaving any extra slot as a style/system/reference tile. Each thumbnail tile must use the same aspect ratio as the requested final slide output. For example, if the final deck is 16:9, every tiny slide in the thumbnail board should be composed as 16:9; if the user requests 3:4 vertical slides, every tiny slide in the thumbnail board should be composed as 3:4.
  40. - Generation gate: before calling Image2.5 for a multi-page deck, always produce `outline.md` and `prompts.md` first, then ask the user to confirm or revise. `prompts.md` must include the selected style's full `Style Lock`.
  41. - Final packaging gate: after Image2.5 generation, show or list the generated images and wait for the user to confirm that all images are approved. Only after that approval, assemble the image-only PPTX, collect all slide images plus `prompts.md`, `outline.md`, and style/reference docs, create a `.zip` package, and save it to the user's default save folder.
  42. ## Workflow
  43. 1. Classify the request:
  44. - Style extraction from images, screenshots, decks, webpages, or descriptions.
  45. - Style library operation: save, list, rename, update, or call a style.
  46. - Document-to-PPT or document-to-image creation.
  47. - Image-only PPT revision or visual redesign with a saved style.
  48. - Post-generation revision of one slide, one image, or one section.
  49. 2. If extracting a style, produce a concise style keyword profile and ask whether to save it as a Markdown file under `styles/`. Save style assets as style files, not as `SKILL.md`.
  50. 3. If applying a style, select the saved style:
  51. - Use explicit names first.
  52. - Treat "终端风格", "CodeBuddy 那种风格", and "科技杂志风格" as `terminal-tech-magazine`.
  53. - If multiple plausible styles exist, ask one short clarifying question.
  54. - Every saved style must provide its own `Style Lock`. Use that style-specific lock as the shared prefix for all Image2.5 prompts in the same deck.
  55. - Style isolation is mandatory: after selecting a style, ignore visual DNA from earlier user-provided references and previous deck attempts unless the user explicitly asks to combine styles.
  56. 4. If creating from a document path, read the document, extract the story arc, audience, page count suggestion, and visual moments before creating prompts or slides.
  57. 5. Before generating images or PPTs, confirm only missing production parameters that materially affect the result: page count, ratio, non-Chinese language requirements, text density, whether the user wants visible dates when the source calls for them, and whether the user wants image files only or an image-only PPTX package. Do not ask whether PPTX must be editable. If the user does not specify language, default to Chinese-first, especially Chinese slide titles. Do not include logos, brand wordmarks, repeated deck identity labels, or watermarks in Image2.5 prompts unless the user explicitly asks; if they ask, recommend post-production placement instead of model-rendered logos.
  58. 6. Do not add date/time by default. If the output must contain any date, time, "today", publication date, log date, report date, or timestamp because the user or source explicitly requires it, verify the date first using the source, conversation environment, local system date, or a reliable time source. Do not invent dates.
  59. 7. Select page types from `references/page-types.md` and map each slide to one page type.
  60. 8. Generate `outline.md` and `prompts.md` before any Image2.5 generation. `outline.md` contains the slide plan and page roles. `prompts.md` contains the full style-specific `Style Lock` plus every page prompt. For multi-page decks, `prompts.md` should also include a thumbnail-board prompt before the per-slide prompts unless the user asks to skip it. Page titles and visible text should be Chinese-first by default, with English retained only for names or terms that should not be translated. Stop and ask the user to confirm these two documents before generating images, unless the user explicitly says to skip confirmation.
  61. 9. For multi-page visual slide output, use this preferred two-pass generation sequence when feasible:
  62. - Pass A: generate one thumbnail board/contact sheet containing all planned slides as small thumbnails. The thumbnail count is flexible but should prioritize 4, 6, 9, or 16 slots. Use the smallest stable slot count that fits the slide count; for 8 slides, use 9 slots and make the extra tile a style/system/reference tile. The board should show only enough readable text to identify each slide and should prioritize layout, color, rhythm, image placement, grid structure, and overall visual consistency.
  63. - The thumbnail tile aspect ratio must match the final requested output ratio. Default final output is 16:9, so default thumbnail tiles are 16:9. If the user asks for 3:4, 4:5, 1:1, or another output ratio, compose every thumbnail tile in that same ratio before generating full-size slides.
  64. - Review the thumbnail board with the user when possible. If the user already said to proceed, use it directly as the visual continuity reference.
  65. - Pass B: generate every slide one by one as a separate full-size image file in the requested final output ratio. Each Image2.5 call or output unit must target exactly one final slide. Include the approved thumbnail board, the selected `Style Lock`, and that slide's page-specific prompt as guidance.
  66. 10. For visual slide output, create full-slide Image2.5 prompts and generate complete slide images only after the user confirms `outline.md` and `prompts.md`. The image itself may contain the final short slide text, but normal inner slides must stay sparse: one main title, one short subtitle or takeaway, and at most 2-3 short points. Each prompt must include the selected style's `Style Lock` so fonts, hierarchy, layout grid, ornaments, color ratios, footer behavior, border/container system, and avoid rules stay consistent across the deck. Separate cover and inner-slide behavior: the cover can carry a strong visual concept; inner slides should prioritize text and information hierarchy, using images and background as quiet support.
  67. 11. Final slide image output is always page-by-page: save and present each slide as its own image file using a stable ordered naming pattern such as `slide-01.png`, `slide-02.png`, etc. Do not rely on a single combined image as the deliverable for a PPT deck. If Image2.5 returns a combined image, treat it only as a draft/reference and regenerate or crop into approved single-page images before PPT assembly.
  68. 12. After images are generated, present the image set for review and ask the user to confirm whether all images are approved or which pages need revision. Do not create the final zip before this approval.
  69. 13. For revision requests, identify the target slide/image, preserve unchanged pages, and update only the affected asset. Repeat image review until the user confirms all images.
  70. 14. After the user confirms all images, assemble each generated image as a full-bleed slide in an image-only PPTX. The PPTX is a delivery container, not an editable deck.
  71. 15. Final package: collect all final images, `deck.pptx`, `prompts.md`, `outline.md`, `style-used.md` if present, the thumbnail board if used, and any revision log into a single `.zip`. Save the zip to the user's default save folder; on macOS, use `~/Downloads` unless the user specified another default folder or the environment clearly provides one.
  72. 16. After generating the prompt set, ask whether the user wants to save the style as a reusable template under `styles/`. If the user says yes, save the style file with its `Style Lock` so future decks can start by choosing this saved style instead of extracting from images again.
  73. 17. Run a final quality check: style consistency, readable Chinese, clean alignment, no overcrowding, unified border/container system, no black outer canvas border, no mixed historical style, no fake date, no model-rendered logo/wordmark drift, one-image-per-slide output, and output parameters respected.
  74. ## Invocation Examples
  75. - "调用终端风格,把这篇文章做成 PPT"
  76. - "用 CodeBuddy 那种风格,把这个文档做成 PPT"
  77. - "提炼这张图的风格,保存成一个可复用风格"
  78. - "列出现在可用的 PPT 风格"
  79. - "用保存的 terminal-tech-magazine 风格生成 10 页 PPT"
  80. - "把这份 Markdown 做成 8 页图片版 PPT,16:9,Image2.5"
  81. - "先给我 outline 和 prompts,我确认后再生成图片"
  82. ## Language Rule
  83. Default to Chinese-first decks and Chinese-first support documents. Titles are especially important: unless the user explicitly requests English or bilingual output, write slide titles, section headers, module labels, and conclusion lines in Chinese. Generated documents such as `outline.md`, `prompts.md`, `style-used.md`, `revision-log.md`, and final delivery notes should also be written mainly in Chinese. Preserve English only for proper nouns, product names, model names, code/API terms, brand names, quoted source terms, and compact labels where translation would reduce clarity, such as `Claude`, `Image2.5`, `API`, `Opus 4.6`, or XML tag names. For bilingual decks, Chinese should usually lead and English should be secondary.
  84. ## Date Rule
  85. Dates and times are off by default. Do not add visible dates, times, timestamps, "today", export time, generated time, or current date unless the user explicitly asks for them, the source article explicitly requires them, or the slide topic itself is date-sensitive. A source file's export timestamp, filename date, filesystem modified time, or the local current date should not appear in the deck unless it is content-relevant.
  86. When date/time is required, verify it before use. If the source provides the required date, use the source date. If the date is relative or operational and the environment date is available, verify with the conversation environment, local system date, or a reliable time source. Do not invent dates. The preferred Chinese date format is `YYYY年M月D日`; the preferred technical log format is `YYYY-MM-DD`.
  87. ## Replaceable Styles
  88. The reusable style library lives in `styles/`. Add new styles as individual Markdown files with stable names such as:
  89. - `terminal-tech-magazine.md`
  90. - `warm-editorial-report.md`
  91. - `minimal-founder-deck.md`
  92. Do not overwrite this skill's workflow to change a visual style. Update or add a style file instead.
  93. Each style file must include a `## Style Lock` section. This section is the cross-page control system for that style and should specify:
  94. - Font family mood and title/body hierarchy.
  95. - Language behavior: Chinese-first visible text and Chinese-first support documents, especially titles, with rules for retaining English terms and product names.
  96. - Relative font scale rules, not exact pixels.
  97. - Layout grid, safe margins, page rhythm, and repeated page structures.
  98. - Reusable decorative elements and how often to use them.
  99. - Style isolation rules: what must not be imported from previous references, previous tasks, or unrelated saved styles.
  100. - Specific color palette with hex values for background, primary, secondary, accent, text, muted text, border, and shadow/glow colors, plus approximate color ratios.
  101. - Border and container rules: one consistent stroke color, stroke thickness mood, corner radius style, card/panel material, divider shape, and line geometry reused across the deck.
  102. - Footer/header metadata behavior.
  103. - Logo/wordmark behavior: default to no generated logos or repeated brand marks; if needed, reserve a quiet empty area for a later fixed asset rather than asking Image2.5 to draw the logo.
  104. - Text density limits: strict maximum visible text for cover pages and inner slides; normal inner slides should be low-density, not report-dense.
  105. - Cover vs inner-slide rules: how expressive the cover can be, and how restrained inner pages must be.
  106. - Image2.5 negative constraints.
README.md · README · 391 lines

View this file on GitHub (read at the repository's current version: the scan did not record which commit it read)

  1. # visual-style-ppt
  2. Style-driven Codex Skill for turning documents, articles, and visual references into image-first presentation decks.
  3. ## Highest Priority Rule
  4. This skill must use the Image 2.5 model for every generated slide image, thumbnail board, visual page, infographic, article illustration, and PPT page image. HTML output is forbidden: do not use HTML/CSS, browser screenshots, SVG/HTML mockups, canvas renders, PIL, or local scripts as final or intermediate image outputs. If Image 2.5 is unavailable, stop and state that Image 2.5 access is required.
  5. <details open>
  6. <summary>中文</summary>
  7. 一个用于「风格驱动 PPT」的 Codex Skill。
  8. 它把视觉风格当成可复用资产:先提炼或选择一个视觉风格,再基于文档、文章、提纲或主题生成图片版幻灯片提示词,并在需要时把逐页图片组装成 image-only PPTX。
  9. ## 仓库简介
  10. `visual-style-ppt` 适合把内容策划、视觉风格提炼、Image2.5 生图提示词和图片版 PPT 打包流程串成一套稳定 SOP。它不追求可编辑 PPT 排版,而是把每页幻灯片视为一张高完成度视觉图片,再用 PPTX 作为最终交付容器。
  11. 推荐 GitHub About 简介:
  12. ```text
  13. Style-driven Codex Skill for creating image-first PPT decks from documents, articles, and visual references.
  14. ```
  15. 推荐 Topics:
  16. ```text
  17. codex-skill, ppt, presentation, image2.5, visual-style, prompt-engineering, slide-deck, markdown-to-ppt
  18. ```
  19. ## 适合做什么
  20. - 把文章、Markdown、文档内容转成图片版 PPT
  21. - 从截图、参考图、网页、已有 deck 中提炼视觉风格
  22. - 保存、复用、维护一套 PPT 视觉风格库
  23. - 用指定风格生成 `outline.md` 和 `prompts.md`
  24. - 生成逐页幻灯片图片,并最终打包成 PPTX 和 zip
  25. - 做小红书信息图、文章配图、视觉化报告页等风格统一的页面资产
  26. ## 核心原则
  27. - **Image 2.5 only**:所有幻灯片图片、缩略图板、信息图和配图都必须由 Image 2.5 模型生成,严禁 HTML 出图或任何等价替代路线。
  28. - **一页一图**:每一页最终幻灯片都应保存为单独图片,例如 `slide-01.png`、`slide-02.png`。
  29. - **风格隔离**:一个 deck 只使用一个选定风格和一个 `Style Lock`,避免混入历史任务或无关参考图的视觉 DNA。
  30. - **低信息密度**:默认内页保持克制,一页一个主标题、一个简短 takeaway,最多 2-3 个信息点。
  31. - **中文优先**:默认使用中文标题、中文模块名和中文支持文档;产品名、模型名、API 等保留英文。
  32. - **不自动加日期**:除非来源或用户明确要求,否则不在页面、脚注或元数据中添加日期和时间。
  33. ## 目录结构
  34. ```text
  35. .
  36. ├── SKILL.md
  37. ├── agents/
  38. │ └── openai.yaml
  39. ├── references/
  40. │ ├── output-package.md
  41. │ ├── page-types.md
  42. │ ├── revision-workflow.md
  43. │ ├── style-interview.md
  44. │ └── workflow.md
  45. └── styles/
  46. ├── french-editorial-commerce.md
  47. ├── impact-grid-editorial.md
  48. └── terminal-tech-magazine.md
  49. ```
  50. ## 安装
  51. 把这个仓库放到 Codex 的 skills 目录下:
  52. ```bash
  53. cd ~/.codex/skills
  54. git clone https://github.com/irenerachel/visual-style-ppt-skill.git visual-style-ppt
  55. ```
  56. 如果你已经有本地目录,也可以直接更新:
  57. ```bash
  58. cd ~/.codex/skills/visual-style-ppt
  59. git pull
  60. ```
  61. ## 使用前提
  62. - 已安装并可使用 Codex Skills。
  63. - 需要生成图片时,当前环境必须具备可调用的 Image 2.5 模型;没有 Image 2.5 时不得用 HTML 或其它图像生成/本地渲染能力替代。
  64. - 需要组装 PPTX 时,环境应具备 PPTX 创建能力;本 Skill 默认生成 image-only PPTX。
  65. - 如果输入是 PDF、Word、Markdown 或网页内容,先确保 Codex 能读取对应文件或链接。
  66. ## 使用示例
  67. ```text
  68. 调用终端风格,把这篇文章做成 8 页 PPT
  69. ```
  70. ```text
  71. 用 CodeBuddy 那种科技杂志风格,把这个 Markdown 做成图片版 PPT,16:9
  72. ```
  73. ```text
  74. 提炼这张图的风格,保存成一个可复用 PPT 风格
  75. ```
  76. ```text
  77. 列出现在可用的 PPT 风格
  78. ```
  79. ```text
  80. 用保存的 terminal-tech-magazine 风格,先给我 outline 和 prompts,我确认后再生成图片
  81. ```
  82. ## 什么时候不适合用
  83. - 需要高度可编辑的 PPT 源文件,例如每个文本框、图表和形状都要可单独修改。
  84. - 需要严格品牌 Logo 复现,并要求模型在每页图片中稳定绘制相同 Logo。
  85. - 需要大量表格、长段落、报告正文堆叠在单页里。
  86. - 需要自动加入今天日期、导出时间或时间戳,但来源没有明确要求。
  87. ## 标准工作流
  88. 1. 判断任务类型:提炼风格、调用风格、文档转 PPT、图片版 PPT 修订,或风格库管理。
  89. 2. 选择或创建一个视觉风格文件。
  90. 3. 从内容中抽取故事线、受众、页数建议和关键视觉时刻。
  91. 4. 先生成 `outline.md` 和 `prompts.md`。
  92. 5. 多页 deck 优先生成 thumbnail board,用于锁定整体节奏。
  93. 6. 用户确认后,再逐页生成最终幻灯片图片。
  94. 7. 用户确认全部图片后,组装 image-only PPTX。
  95. 8. 打包最终图片、PPTX、提示词、提纲、风格文件和修订记录。
  96. ## 风格库
  97. 可复用风格放在 `styles/` 目录,每个风格是一个 Markdown 文件。
  98. 每个风格文件都应该包含 `## Style Lock`,用于控制整套 deck 的字体气质、布局网格、色彩比例、边框系统、页眉页脚、文本密度、Logo 处理和 Image2.5 negative constraints。
  99. 当前内置风格:
  100. | 风格文件 | 适合内容 | 视觉气质 |
  101. | --- | --- | --- |
  102. | `terminal-tech-magazine.md` | AI、开发者工具、技术产品、深色视觉报告 | 终端科技杂志风格 |
  103. | `impact-grid-editorial.md` | 观点型内容、趋势判断、强标题页面 | 冲击力网格编辑风格 |
  104. | `french-editorial-commerce.md` | 品牌、消费、生活方式、审美型商业内容 | 法式编辑商业风格 |
  105. ## 新增风格规范
  106. 新增风格时,在 `styles/` 下创建一个稳定命名的 Markdown 文件,例如:
  107. ```text
  108. styles/minimal-founder-deck.md
  109. ```
  110. 每个风格至少包含:
  111. - 风格定位和适用场景
  112. - 颜色系统,最好包含 hex 色值
  113. - 字体气质与标题/正文层级
  114. - 页面网格、留白、页眉页脚规则
  115. - 卡片、边框、分割线、装饰元素规则
  116. - `## Style Lock`
  117. - Image2.5 negative constraints
  118. ## 生成物约定
  119. 推荐输出文件包括:
  120. - `outline.md`
  121. - `prompts.md`
  122. - `style-used.md`
  123. - `thumbnail-board.png`
  124. - `slide-01.png`
  125. - `slide-02.png`
  126. - `deck.pptx`
  127. - `revision-log.md`
  128. - final `.zip`
  129. PPTX 是交付容器,不是可编辑排版源文件。真正的视觉结果以逐页图片为准。
  130. ## 命名约定
  131. - 风格文件使用 kebab-case,例如 `terminal-tech-magazine.md`。
  132. - 生成图片使用两位数序号,例如 `slide-01.png`。
  133. - 中间文档使用稳定文件名,例如 `outline.md`、`prompts.md`、`revision-log.md`。
  134. - 同一个项目的最终交付建议放入独立目录,避免不同 deck 的图片混在一起。
  135. ## 维护建议
  136. - 修改工作流时更新 `SKILL.md` 和 `references/`。
  137. - 新增视觉风格时只添加 `styles/*.md`,不要为了改风格重写 `SKILL.md`。
  138. - 每个风格都要写清楚自己的 `Style Lock`。
  139. - 避免在风格文件里写过多一次性项目内容,保持它可复用。
  140. - 如果生成流程、包装方式或修订规则变化,同步更新 `references/` 中对应文档。
  141. ## 授权与使用
  142. 当前仓库按个人 Skill 仓库维护。若后续计划公开分发、接受外部贡献或作为模板复用,建议再补充正式开源协议,例如 MIT License。
  143. </details>
  144. <details>
  145. <summary>English</summary>
  146. A Codex Skill for style-driven visual presentation creation.
  147. This skill treats visual style as a reusable asset: extract or select a visual style first, then use documents, articles, outlines, or topics to create image-based slide prompts. When needed, it assembles the generated slide images into an image-only PPTX deck.
  148. ## Repository Summary
  149. `visual-style-ppt` connects content planning, visual style extraction, Image2.5 prompt writing, and image-first PPT packaging into a reusable workflow. It does not aim to create fully editable PowerPoint layouts. Instead, it treats each slide as a polished visual image and uses PPTX as the final delivery container.
  150. Suggested GitHub About description:
  151. ```text
  152. Style-driven Codex Skill for creating image-first PPT decks from documents, articles, and visual references.
  153. ```
  154. Suggested topics:
  155. ```text
  156. codex-skill, ppt, presentation, image2.5, visual-style, prompt-engineering, slide-deck, markdown-to-ppt
  157. ```
  158. ## What It Is For
  159. - Turn articles, Markdown files, and documents into image-based presentations
  160. - Extract visual style from screenshots, reference images, webpages, or existing decks
  161. - Save, reuse, and maintain a visual style library for PPT creation
  162. - Generate `outline.md` and `prompts.md` from a selected style
  163. - Generate one slide image per page and package the result as PPTX and zip
  164. - Create Xiaohongshu-style infographics, article visuals, visual reports, and consistent page assets
  165. ## Core Principles
  166. - **Image 2.5 only**: Every slide image, thumbnail board, infographic, and illustration must be generated with the Image 2.5 model; HTML image output and equivalent substitute routes are forbidden.
  167. - **One image per slide**: Each final slide should be saved as an individual image, such as `slide-01.png` or `slide-02.png`.
  168. - **Style isolation**: Each deck uses exactly one selected style and one `Style Lock`, avoiding visual DNA from past tasks or unrelated references.
  169. - **Low text density**: Inner slides stay calm by default: one title, one short takeaway, and at most 2-3 information points.
  170. - **Chinese-first by default**: Slide titles, section labels, and support documents default to Chinese; product names, model names, and API terms stay in English when clearer.
  171. - **No automatic dates**: Dates and timestamps are not added unless the source or the user explicitly requires them.
  172. ## Structure
  173. ```text
  174. .
  175. ├── SKILL.md
  176. ├── agents/
  177. │ └── openai.yaml
  178. ├── references/
  179. │ ├── output-package.md
  180. │ ├── page-types.md
  181. │ ├── revision-workflow.md
  182. │ ├── style-interview.md
  183. │ └── workflow.md
  184. └── styles/
  185. ├── french-editorial-commerce.md
  186. ├── impact-grid-editorial.md
  187. └── terminal-tech-magazine.md
  188. ```
  189. ## Installation
  190. Place this repository under your Codex skills directory:
  191. ```bash
  192. cd ~/.codex/skills
  193. git clone https://github.com/irenerachel/visual-style-ppt-skill.git visual-style-ppt
  194. ```
  195. If you already have a local copy, update it with:
  196. ```bash
  197. cd ~/.codex/skills/visual-style-ppt
  198. git pull
  199. ```
  200. ## Requirements
  201. - Codex Skills should be installed and available.
  202. - Image generation requires the Image 2.5 model in the current environment; HTML and other image-generation or local-rendering substitutes are not allowed.
  203. - PPTX assembly requires an environment capable of creating PPTX files; this skill defaults to image-only PPTX.
  204. - If the input is a PDF, Word document, Markdown file, or webpage, make sure Codex can access the file or link.
  205. ## Example Prompts
  206. ```text
  207. Use the terminal-tech style and turn this article into an 8-slide PPT.
  208. ```
  209. ```text
  210. Use the CodeBuddy-like tech magazine style to turn this Markdown file into a 16:9 image-based presentation.
  211. ```
  212. ```text
  213. Extract the style from this image and save it as a reusable PPT style.
  214. ```
  215. ```text
  216. List the available PPT styles.
  217. ```
  218. ```text
  219. Use the saved terminal-tech-magazine style. Give me the outline and prompts first, then wait for confirmation before generating images.
  220. ```
  221. ## When Not To Use It
  222. - You need a highly editable PPT source file where every text box, chart, and shape must remain separately editable.
  223. - You need strict logo reproduction and expect the image model to redraw the same logo consistently on every page.
  224. - You need dense tables, long paragraphs, or report-style body copy packed into a single slide.
  225. - You need automatic dates, export times, or timestamps when the source does not explicitly require them.
  226. ## Standard Workflow
  227. 1. Classify the task: style extraction, style application, document-to-PPT, image-only PPT revision, or style library management.
  228. 2. Select or create a visual style file.
  229. 3. Extract the story arc, audience, page count suggestion, and key visual moments from the source.
  230. 4. Generate `outline.md` and `prompts.md` first.
  231. 5. For multi-page decks, generate a thumbnail board first to lock the visual rhythm.
  232. 6. After user confirmation, generate each final slide image one by one.
  233. 7. After all images are approved, assemble an image-only PPTX.
  234. 8. Package the final images, PPTX, prompts, outline, style file, and revision log.
  235. ## Style Library
  236. Reusable styles live in the `styles/` directory. Each style is a Markdown file.
  237. Every style file should include a `## Style Lock` section that controls the deck's font mood, layout grid, color ratios, border system, headers and footers, text density, logo handling, and Image2.5 negative constraints.
  238. Built-in styles:
  239. | Style file | Best for | Visual mood |
  240. | --- | --- | --- |
  241. | `terminal-tech-magazine.md` | AI, developer tools, technical products, dark visual reports | Terminal tech magazine |
  242. | `impact-grid-editorial.md` | Opinion pieces, trend analysis, strong title pages | High-impact editorial grid |
  243. | `french-editorial-commerce.md` | Brand, consumer, lifestyle, aesthetic business content | French editorial commerce |
  244. ## Adding A New Style
  245. Create a stable Markdown file under `styles/`, for example:
  246. ```text
  247. styles/minimal-founder-deck.md
  248. ```
  249. Each style should include:
  250. - Style positioning and suitable use cases
  251. - Color system, preferably with hex values
  252. - Font mood and title/body hierarchy
  253. - Page grid, spacing, header, and footer rules
  254. - Card, border, divider, and decorative element rules
  255. - `## Style Lock`
  256. - Image2.5 negative constraints
  257. ## Output Convention
  258. Recommended output files include:
  259. - `outline.md`
  260. - `prompts.md`
  261. - `style-used.md`
  262. - `thumbnail-board.png`
  263. - `slide-01.png`
  264. - `slide-02.png`
  265. - `deck.pptx`
  266. - `revision-log.md`
  267. - final `.zip`
  268. The PPTX is a delivery container, not the editable layout source. The final visual result is defined by the per-slide images.
  269. ## Naming Conventions
  270. - Use kebab-case for style files, such as `terminal-tech-magazine.md`.
  271. - Use two-digit numbering for generated images, such as `slide-01.png`.
  272. - Use stable names for intermediate documents, such as `outline.md`, `prompts.md`, and `revision-log.md`.
  273. - Keep each final project in its own output directory to avoid mixing images from different decks.
  274. ## Maintenance
  275. - Update `SKILL.md` and `references/` when changing the workflow.
  276. - Add new visual styles as `styles/*.md`; do not rewrite `SKILL.md` just to change a style.
  277. - Make sure every style includes a clear `Style Lock`.
  278. - Keep style files reusable and avoid embedding too much one-off project content.
  279. - If the generation flow, packaging rules, or revision workflow changes, update the matching document in `references/`.
  280. ## License And Usage
  281. This repository is currently maintained as a personal skill repository. If you plan to distribute it publicly, accept external contributions, or use it as a reusable template, consider adding a formal open source license such as the MIT License.
  282. </details>

Ratings and comments

Sign in to rate this skill.

Sign in

Free to use. If it’s helped, pass it on.