SkillFoxx

SlideMuse skill

SlideMuse turns a PDF, paper or notes into an image-based deck in 15 styles and can rebuild it as an editable PPTX

  • 215 ★ on the repository
  • Last commit
  • MIT
Skill

Extended access

Why this level

  • The installer sets up isolated dependencies and the skill reads supplied documents and writes PPTX and image files
All reasons and checks

helloo1568/slidemuseAuthor:helloo1568

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add skills/slidemuse

Detects the agents on your machine, shows the access level and pins the version.

Other ways to install

Access and checks

How we review

Extended accessThe tool runs code or makes network calls and only reads external data.

  • The installer sets up isolated dependencies and the skill reads supplied documents and writes PPTX and image files

This is third-party code. Review the repository files before installing.

What it does

SlideMuse is a skill for Codex, Claude Code, OpenCode, WorkBuddy, Trae and Qoder that carries a presentation from source material to a file. The agent first drafts a slide-by-slide content plan and shows four visual style options to pick from. After the choice it generates the full set of slide images and assembles a PPTX. There are 15 ready styles with two slides each. The second stage is optional: from the images the skill rebuilds an editable PPTX. Text, shapes and tables become native PowerPoint objects, while complex illustrations and photos split into separate replaceable layers. Inputs can be finished slide images, a scanned PDF or an image-based PPTX. The skill supports speaker notes, task recovery and local revisions. A separate section targets competition defenses: innovation and entrepreneurship, technical contests, research and teaching presentations.

Who it is for. Students, teachers and managers who need a finished deck from a report or paper in one style, with the option to keep editing the file in PowerPoint.

Good fit when

  • A PDF report has to become a 10-slide presentation for a team
  • A competition defense needs a deck built around problem, solution, validation and value
  • You have slide images or a scan that must become an editable PPTX
  • Chart figures change and the slides need rebuilding from new data

Not a fit when

  • You need a deck strictly from a company template: the skill draws its own styles
  • Your agent has no image generation: it is taken from the agent itself
  • You want a quick text draft with no style choice or reconstruction

Example request

Use the slidemuse skill to turn this report into a 10-slide team deck, showing the plan and four styles first

Limitations

  • The skill relies on the agent's abilities: document reading, vision, image generation and local file work. Without them the stages do not run.
  • PPTX rendering needs an installed PowerPoint or LibreOffice with Poppler. Python 3.10 or newer and Git for cloning are required.
  • Showcase examples in the README are Chinese. English versions of the text exist, but the showcase slide content is Chinese.
  • Editable reconstruction rebuilds objects from images, and complex illustrations stay as image layers with no vector conversion.

How to disable. Delete the `slidemuse` skill folder from the client skills directory and the `slidemuse` directory with isolated dependencies that `install.py` created.

FAQ

How do I install SlideMuse?

Clone the repository and run `python slidemuse/install.py`. If several clients are installed, add `--client codex`, `workbuddy`, `trae` or `qoder`. Chinese editions use `trae-cn` and `qoder-cn`.

What is needed to render PPTX?

An installed PowerPoint or LibreOffice with Poppler. Document reading, vision and image generation come from the agent itself.

Can I edit the finished deck?

Yes, at the second stage. The skill builds an editable PPTX: text, shapes and tables become PowerPoint objects and photos and complex drawings become separate layers.

How many styles are included?

Fifteen, two slides per style. Before generation the agent shows four options and you pick one.

Is there a mode for competition defenses?

Yes. The skill builds a deck around the problem, solution, innovation, validation and value and lists scenarios: innovation, technical contests, research and teaching.

Discussion

No questions or reviews yet.

Skills

Image to Editable PPT

Image to Editable PPT Skill

Needs a VPN

A Codex skill that rebuilds slides from images, PDFs and image-based PPTX files into an editable PowerPoint deck

SkillExtended access3.3K

A Claude Code and Codex workflow that makes native PPTX decks where every element stays editable in PowerPoint.

WorkflowMinimal access313

Skill turns the agent into a deck producer: brief, storyboard, slide posters, QA and a PDF or editable PPTX.

SkillMinimal access54

The image-to-ppt skill turns images, PDFs and picture-only slides into a PowerPoint deck whose text and elements can be edited.

PluginMinimal access62

A Codex skill that rebuilds a slide from an image or PDF into PowerPoint where text, tables and arrows stay editable objects.

SkillExtended access141

A skill that builds a PowerPoint deck with live text, charts and tables: an image model draws each slide, the agent rebuilds it.

SkillExtended access99

README in short

The README presents SlideMuse as a way to make attractive yet editable presentations. It shows sample work and the reconstruction idea, then competition scenarios, a quick start with a ready phrase for the agent, the path from material to file and links to guides, references and the community.

README badge

Original SKILL.md

---
name: slidemuse
description: 将书籍、PDF、论文、报告或文字制作成视觉统一的图片版 PPT，适用于挑战杯、创新大赛、计算机设计、软件杯、服创、信息安全、集成电路、嵌入式、机械创新、化工设计、结构设计、智能车外赛项、研创赛等学科竞赛答辩，MPA/MPAcc 等案例陈述，教创赛与师范生说课，以及路演、答辩、课程展示和读书分享；按需还原可编辑 PPTX，也支持直接还原已有图片、PDF 或图片版 PPTX。不用于普通修改已有可编辑 PPTX。
---

# SlideMuse · 2.25.1

新材料按 **内容提炼 → 图片 PPT → 按需可编辑还原** 制作。已有页面要求还原时直接进入重建。内容与视觉应准确、清楚、统一；检查服务于这几个结果。

客户端调用：Codex 可用 `$slidemuse`，Claude Code / Qoder 可用 `/slidemuse`；WorkBuddy / Trae 可在技能面板选择，或明确说“使用 slidemuse 技能”。各客户端共用本工作流与本地 Python 工具，图像生成与编辑使用当前宿主实际提供的能力。安装位置和国内版选项见[客户端适配](docs/clients.md)。

## 开始任务

先检查用户已给的材料和上下文，必要时阅读材料来判断篇幅与用途。只询问会实质改变结果、又无法合理推断的信息，并一次问完；不要为了填满五个字段阻止阅读。

- 源材料不可访问、受众/任务存在相互冲突的解释、关键事实缺失时询问。
- 未指定风格：按无参考处理；未指定页数：根据材料推荐并随内容稿展示；未指定交付：先做图片版；默认 16:9、沿用材料语言。使用场景可从上下文判断，无法判断且影响内容取舍时再问。把采用的默认值简要告诉用户，不额外等待确认默认值。
- 用户已要求可编辑版，即已授权最后的重建阶段，不在图片版完成后重复询问。
- 默认只在**内容方案**和**风格选择**处请用户决策。已有“你来定”“按这个做”“其余由你决定”等对应授权时继续执行，记录授权范围即可。明确修改也是该范围的授权。
- 状态码、哈希、日志与检查文件由 Agent 管理，不要求用户理解或填写。进度只说明当前成果、需要选择的内容和实际阻塞项。

任务文件放到独立工作目录。用 `deck-spec.md` 保存内容稿、采用的需求、用户授权和当前进度；[规格模板](references/deck-spec-template.md)可按任务规模选用，保留实际信息，不机械填满无关栏目。恢复旧任务或修改已有交付时再读[变更与恢复](references/workflow-updates.md)。

## 1. 内容方案

读取 [Prompt 1A](references/prompts.md)、[内容深度](references/content-outline.md)与[密度规则](references/content-density.md)。论文、政策或长报告再按需读[材料审阅](references/material-reading.md)。

先明确受众需要理解的问题和整套回答，再写逐页内容：页面作用、结论、实际展示文案、证据/数据及必要解释。内容页按其用途安排信息密度；图表能表达的关系不重复堆正文。封面和过渡页按作用简化。

核对材料事实、推断、计算和建议，修正无支撑结论、错误口径、缺失关键解释及篇幅问题。来源不够时缩小结论或询问必要事实，不编造数据。地图、实验截图等证据按[证据素材](references/evidence-assets.md)决定保留原图或可靠数据重绘。

向用户展示整套叙事和逐页实际内容；长稿可分段呈现，不能只给标题表或文件链接。没有内容授权时在这里等待；已有对应授权时展示后继续。内容确定后再创建 Page Spec、风格图片和成品，不提前消耗生图调用。

## 2. 风格和图片版

读取 [Prompt 1B / 2](references/prompts.md)与[风格总览](references/style-options.md)。默认给四套使用相同内容的独立缩略图总览，让用户选择。沿用用户明确指定的模板，或按其明确的跳过/代选授权执行；已有选择不再询问。代选时仍生成四套并记录选择理由，跳过预览时记录 `skipped`。

总览使用任务指定页面比例；最多八页时展示全部，更多页时选同一组六至八个代表页。每套一次生图，保留方案编号。纯尺寸、网格或边距问题用 `build_style_overview.py` 修正；内容或视觉问题才重试生图。实际生成四套时运行 `validate_style_options.py` 并看图核对，只展示每套最新有效候选。

内容与风格确定后，按 [Page Spec](references/page-spec.md) 创建 `page-spec.json`，保留准确文案、数据、来源、稳定 ID、风格和授权。已有用户授权可直接记录为内容确认，不另索要固定口令。

```sh
python "<skill-dir>/slidemuse.py" validate "<work>/page-spec.json" --strict
```

Report a problem