Sci Diagram Pptx logo

Sci Diagram Pptx

Organization
studio-yebisu
sci-diagram-pptx

Reconstruct or repair user-provided scientific and academic schematics as visually faithful, native editable PowerPoint (.pptx) files with executable build source; also inspect existing PPTX files for meaningful fidelity and editability defects. Use for 复刻、还原、临摹、修复或检查科研框架图、技术路线图、科研流程图、机制图、算法流程图、科研系统结构图、学术概念模型、结构化科研信息图,以及包含数学符号或公式的学术图示. Do not use for quantitative data visualizations or statistical plots whose meaning is encoded by axes, scales, legends, or data-driven geometry; general business diagrams; organization charts; ordinary deck design; OCR-only extraction; or simple image placement.

Overview

Publisherstudio-yebisu
Repositoryyaoyorozu-ai-toolbox-slides
Skill namesci-diagram-pptx
Stars
10
Forks
2
Bundled files
15
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 15 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by studio-yebisu on GitHub. Read the source before you install it.

Installation

Install the Sci Diagram Pptx AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/studio-yebisu/yaoyorozu-ai-toolbox-slides/tree/main/examples/repositories/sci-diagram-reconstruction/upstream
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/studio-yebisu/yaoyorozu-ai-toolbox-slides.git /tmp/yaoyorozu-ai-toolbox-slides
mkdir -p .claude/skills
cp -r /tmp/yaoyorozu-ai-toolbox-slides/examples/repositories/sci-diagram-reconstruction/upstream .claude/skills/sci-diagram-pptx
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Sci Diagram Pptx in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Sci Diagram Pptx on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Sci Diagram Pptx is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

SciDiagram PPTX

Reproduce the supplied scientific schematic; do not redesign it unless the user explicitly asks.

Preserve what matters

Prioritize, in order:

  1. visible wording, symbols, formulas, and scientific meaning;
  2. nodes, regions, nesting, reading order, and connector topology;
  3. native editability of the meaning-bearing structure;
  4. readable layout and valid export.

Treat small differences in antialiasing, kerning, exact color, line weight, corner radius, arrowhead proportions, or local spacing as cosmetic when meaning and readability remain sound. Do not chase pixel identity.

Use this skill when meaning is carried mainly by boxes, labels, formulas, nested regions, arrows, or explicit relationships. Do not use it for charts whose evidence is encoded by axes, scales, legends, or data-driven geometry. Split a mixed figure by origin and reconstruct only the selected schematic panel. Ask about unreadable content only when a character, formula, or arrow direction could change meaning.

Never silently rewrite wording, simplify topology, recolor the figure, invent a cycle from visual layout, or reverse a relationship because another design appears clearer.

Choose exactly one route

  • Reconstruction: an image or selected image panel becomes a new single-slide editable PPTX. Read reconstruction.md.
  • Repair: named defects in an existing PPTX are corrected without changing unrelated slides. Read repair.md.
  • Inspection: named slides in an existing PPTX are diagnosed without modification. Read inspection.md.

Read only the selected route reference. Do not load the other route contracts. If a request mixes routes, complete the named deliverable and report separate findings briefly.

Select one runtime only when authoring

Reconstruction and Repair choose one host-provided backend before writing build.mjs:

  1. Honor an explicit SCI_DIAGRAM_RUNTIME=artifact-tool or SCI_DIAGRAM_RUNTIME=pptxgenjs value.
  2. When unset for Reconstruction, prefer artifact-tool when the installed OpenAI Presentations runtime is available; otherwise use pptxgenjs only when its locked package and render dependencies are ready.
  3. Repair requires a backend that can import and preserve the existing deck. Use artifact-tool when available. PptxGenJS cannot perform Repair; stop or obtain approval to change the task to standalone Reconstruction.
  4. Probe only when availability is uncertain. Do not repeat a successful probe for the same backend, version, task directory, and host unless the environment changes or a real invocation fails.
  5. Do not switch backend after build.mjs exists or mix both authoring packages in one script.

Read exactly one runtime reference:

Inspection selects no authoring backend.

Shared construction invariants

  • Record relationships from the source, not from proximity or circular placement. A line label is not a node. Preserve origin, destination, direction, dash meaning, crossings, nesting, and z-order.
  • Give every visible element exactly one owner: native or raster. Text or arrows retained inside an inset must not also be redrawn natively. If a native overlay is essential, exclude or cover the raster copy in the initial ownership map rather than discovering duplicates through repeated renders.
  • Keep photographs, microscopy, experimental results, model screenshots, and complex composite mini-figures as replaceable local raster insets when native reconstruction would be dishonest. Keep their semantic frame and external annotations native.
  • Never use a full-slide source bitmap, hidden tracing image, imported path cloud, or image-tile mosaic on the editable slide.
  • Preserve source aspect ratio and use one source-to-slide transform. Keep topology and object data explicit in the executed build source.
  • Ask before a fallback changes scientific meaning, formula content, arrow topology, or converts a substantial meaning-bearing region to raster. Use a close native approximation without pausing when the difference is cosmetic.

Bound the work

For Reconstruction and Repair, the default budget is two authored passes at most:

  1. Build pass-01, then run the selected render and the package checker. Run those two read-only checks concurrently when the host supports it.
  2. Combine checker hard failures with named visual blockers. Warnings and cosmetic differences do not authorize another pass.
  3. If no blocker remains, deliver the accepted pass.
  4. If blockers remain, collect all of them before editing and make one grouped correction in pass-02.
  5. The second render/check is terminal. Deliver or report the named blocker. A third authored pass requires an explicit user request.

Use --pass-index 1 and --pass-index 2 with the checker. Its decision.next_action is deliver, repair_once, or stop_with_blocker. Do not create final, approved, release, or similarly renamed extra passes. Do not use a pixel-difference percentage as a substitute for checking wording, formulas, topology, and visible clipping.

Inspection is read-only and never starts a repair loop.

Read optional references only when triggered

Use at most one focused capability probe for one unresolved feature. Reuse a result already established for the same backend and version.

Portability and handoff

Use portable construction plus the normal render and structure check as the validation boundary. A successful render proves only the selected renderer opened and exported the file; it does not prove pixel identity across applications. Do not create platform variants by default.

For Reconstruction or Repair, return a short note stating the authoring runtime, accepted pass, checks performed, and any material approximation or local raster inset. Inspection returns findings only.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Sci Diagram Pptx AI skill do?

Reconstruct or repair user-provided scientific and academic schematics as visually faithful, native editable PowerPoint (.pptx) files with executable build source; also inspect existing PPTX files for meaningful fidelity and editability defects. Use for 复刻、还原、临摹、修复或检查科研框架图、技术路线图、科研流程图、机制图、算法流程图、科研系统结构图、学术概念模型、结构化科研信息图,以及包含数学符号或公式的学术图示. Do not use for quantitative data visualizations or statistical plots whose meaning is encoded by axes, scales, legends, or data-driven geometry; general business diagrams; organization charts; ordinary deck design; OCR-only extraction; or simple image placement.

Why use Sci Diagram Pptx on TypingMind?

Because you install it once and use it with any model. Sci Diagram Pptx is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Sci Diagram Pptx in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/studio-yebisu/yaoyorozu-ai-toolbox-slides/tree/main/examples/repositories/sci-diagram-reconstruction/upstream. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Sci Diagram Pptx?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Sci Diagram Pptx?

As many as you like. As long as a model supports skills, you can use Sci Diagram Pptx with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Sci Diagram Pptx AI skill free?

It is published on GitHub by studio-yebisu. Check the repository for licensing terms. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇