Academic Slide Pragmatic Fallback logo

Academic Slide Pragmatic Fallback

Community
fangyuanopus
academic-slide-pragmatic-fallback

Create Chinese academic literature-report PPTs from real paper/SI figures. Use for journal-club, thesis-defense, paper-to-PPT, and slide-review tasks, especially when Image2-style full-slide generation is unavailable. Preserve the strict Image2-only route when available; otherwise require explicit consent and choose a clean template-fidelity or a sample-style pragmatic fallback with real figures, source tracking, structural QA, and render QA.

Overview

Publisherfangyuanopus
Repositoryliterature-report-ppt-builder
Skill nameacademic-slide-pragmatic-fallback
Stars
121
Forks
6
Bundled files
68
LicenseMIT
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.

  • 68 bundled files

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

  • Open source

    Published by fangyuanopus on GitHub. Read the source before you install it.

Installation

Install the Academic Slide Pragmatic Fallback 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/fangyuanopus/literature-report-ppt-builder/tree/main/skills/academic-slide-pragmatic-fallback
  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/fangyuanopus/literature-report-ppt-builder.git /tmp/literature-report-ppt-builder
mkdir -p .claude/skills
cp -r /tmp/literature-report-ppt-builder/skills/academic-slide-pragmatic-fallback .claude/skills/academic-slide-pragmatic-fallback
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Academic Slide Pragmatic Fallback 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 Academic Slide Pragmatic Fallback 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 Academic Slide Pragmatic Fallback 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.

Academic Slides with Safe Fallbacks

Create academic decks that are evidence-led, readable, and honest about their production route.

Non-negotiable rules

  • Use scientific visuals only from the paper, SI, or user-provided sources. Never fabricate, redraw, or semantically alter data figures, microscopy, spectra, molecular structures, mechanisms, tables, or charts.
  • Write conclusion-style titles, trace every visual in a figure-source manifest, and calibrate claims to evidence.
  • Do not call an editable or code-built deck Image2-only.

Choose one route

Route A: Image2-only authoritative deck

Use only when a full-slide Image2 backend is confirmed and can faithfully preserve supplied scientific crops. Keep image2_manifest.json, package one approved full-slide image per PPTX slide, and run scripts/validate_image_only_pptx.py.

Route B: template-fidelity fallback

Use only after explicit user consent and a passed hygiene audit. The template owns its navigation, footer, and mapped slots; code edits inherited objects in place and must not add a second navigation layer.

When the user explicitly asks to reproduce the bundled sample template exactly, use Route B's exact-template mode. Preserve the sample's navigation labels and decorative chrome intentionally, map each output page to a distinct source-slide archetype, and replace or remove every source-specific scientific object. Do not describe a code-redrawn Route C deck as an exact reproduction. Read references/exact-template-fidelity.md before building.

text
python scripts/build_fallback_template_pptx.py edit_plan.json exact.pptx --strict --fail-on-warnings
python scripts/audit_fallback_template_pptx.py exact.pptx --build-report exact.build-report.json --check-masters --exact-template --fail-on-review
python scripts/audit_exact_template_fidelity.py exact.pptx --plan edit_plan.json --build-report exact.build-report.json --fail-on-review
text
python scripts/audit_fallback_template_pptx.py template.pptx --check-masters --fail-on-review

Reject templates containing old scientific content, figure labels, sample names, or non-removable master/layout objects. For adaptive cross-domain reuse, also reject fixed legacy navigation. The only exception is exact-template mode, where the user has explicitly requested that navigation and it remains semantically appropriate. Read references/template-hygiene.md and references/fallback-template-pptx.md.

Route C: pragmatic fallback

Use only after explicit user consent when Image2 is unavailable and no clean, domain-compatible template passes Route B. Start from a blank 16:9 presentation: code owns every visible object. Never import a content-bearing master or layout.

Route C follows the visual grammar of assets/sample-literature-report.pptx—white canvas, red-black-gray palette, shallow top navigation, conclusion title, thin rule, and quiet footer/page number—without inheriting the sample's old content. Read references/first-template-style-contract.md and references/pragmatic-fallback.md.

Every page brief must specify one of: cover, text, figure_right, figure_wide, process, comparison, summary, or closing.

text
python scripts/build_pragmatic_fallback_pptx.py pragmatic_edit_plan.json --out editable_fallback_presentation.pptx
python scripts/validate_pragmatic_fallback_pptx.py editable_fallback_presentation.pptx --plan pragmatic_edit_plan.json --fail-on-review

Every displayed scientific figure requires source_type: figure_crop, crop_verified: true, figure ID, source page, and source label. A full paper-page scan is not a figure crop.

Page ownership and fit contract

  • Choose exactly one owner per slide: inherited template objects in Route B, or generated objects in Route C. Never combine the two.
  • The builder treats every code-owned text box as a fixed template slot: it blocks copy that exceeds the slot's CJK-aware line capacity, as well as titles over 54 characters and overlong bullet lists. Shorten or restructure copy before building; never shrink one page's font or truncate with ellipses.
  • Select the layout from the evidence: use wide figures for plots/tables, right-side figures for portrait images, process flow for 2–5 steps, comparison for two alternatives, and summary rows for 2–5 conclusions/parameters.
  • Do not expose internal route names, template notes, model names, fallback labels, or build instructions in audience-facing slides.

Required production flow

  1. Confirm paper/SI source availability and explicit fallback consent.
  2. Create paper_analysis.json following references/paper-analysis-contract.md, then run scripts/validate_paper_analysis.py --fail-on-review. This is the source of truth for the logic tree, figure boundaries, evidence chains, adaptive navigation, and slide claims.
  3. Derive the terminology table, figure-source manifest, deck_order_map.md, and page briefs from the validated analysis. Do not introduce a slide claim, source figure, or navigation section that is absent from the analysis without updating and revalidating it.
  4. Choose Route A, B, or C and prepare page briefs with source crops and page types.
  5. Build only through that route; do not silently substitute a different route.
  6. Run structural QA, render every slide with PowerPoint/LibreOffice where available, inspect individual pages and a montage, and fix all overflow, overlap, legacy residue, unreadable figures, and inconsistent chrome.

For exact-template mode, also compare the montage with the source-template montage. Passing structural QA alone is insufficient: the inherited navigation, title treatment, margins, footer band, image framing, and whitespace rhythm must remain visibly unchanged.

Route C delivery statement

text
Image2 backend used: no
Delivery mode: pragmatic editable fallback accepted by user
Visual style: sample-deck-derived white/red/black academic grammar
Scientific visuals: real paper/SI/user-provided figure crops only
Structural QA completed: yes
Render/layout QA completed: yes/no
Editable elements may exist: yes

Bundled files

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

and 8 more files.

Frequently asked questions

What does the Academic Slide Pragmatic Fallback AI skill do?

Create Chinese academic literature-report PPTs from real paper/SI figures. Use for journal-club, thesis-defense, paper-to-PPT, and slide-review tasks, especially when Image2-style full-slide generation is unavailable. Preserve the strict Image2-only route when available; otherwise require explicit consent and choose a clean template-fidelity or a sample-style pragmatic fallback with real figures, source tracking, structural QA, and render QA.

Why use Academic Slide Pragmatic Fallback on TypingMind?

Because you install it once and use it with any model. Academic Slide Pragmatic Fallback 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 Academic Slide Pragmatic Fallback in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/fangyuanopus/literature-report-ppt-builder/tree/main/skills/academic-slide-pragmatic-fallback. 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 Academic Slide Pragmatic Fallback?

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 Academic Slide Pragmatic Fallback?

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

Is the Academic Slide Pragmatic Fallback AI skill free?

Yes. It is published on GitHub by fangyuanopus under the MIT license. 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 👇