BeClaude

present

New
GitHub TrendingGeneralby OrnlyP63

Transform raw .md files into Beamer LaTeX presentations (PDF). Auto-detects the best presentation framework (Minto Pyramid, SCR, STAR, BLUF, IMRaD, and 17 more) from content signals, restructures content into that framework's narrative flow, and generates a compilable Beamer .tex file. Use when user provides a .md file and asks to make a presentation, create slides, or export to PDF.

Community PluginView Source

Overview

Present

Transform raw .md notes into structured Beamer LaTeX presentations.

Workflow

Step 1 — Read inputs

  1. Read the raw .md file the user provided
  2. Read references/research.md — contains 22 frameworks + auto-detection decision tree

Step 2 — Detect framework

Scan raw .md for detect_signal keywords from each framework in references/research.md. Use the decision tree at the top of references/research.md as the primary guide.

Pick the top match. Also pick a default Beamer theme:

Content typeBeamer theme
Academic / researchAnnArbor
Business / corporateWarsaw
Technical / engineeringCambridgeUS
Creative / storytellingDarmstadt
General / unclearMadrid

Step 3 — Confirm with user

Show exactly this before proceeding:

code
Detected framework: [NAME] — [one-line reason based on signals found]
Beamer theme: [THEME]

Confirm? (y) or override framework/theme:

Wait for user response. Accept override if given.

Step 4 — Restructure content

Rewrite the raw .md content following the chosen framework's structure from references/research.md:

  • Use framework's Structure field as the slide order
  • Apply Transition Phrases between sections
  • Avoid framework Pitfalls
  • Target the framework's Slide Count range

Step 4.5 — Research gap-filling

After restructuring, scan each framework section for two gap types:

  1. Missing section — a slot required by the framework's Structure field has no content from the source
  2. Thin section — a section cannot stand as a coherent slide given the framework's expected content density (semantic + framework-relative: would produce < 2 substantive bullet points)

For each gap found:

  • Formulate targeted WebSearch query: "[topic] [missing section name]"
  • Synthesize results into 3–5 slide-ready bullet points
  • Tag additions with \footnote{\textit{[AI-researched]}} in the .tex

After full research pass, show user:

code
Research found [N] gap(s):
- [Section name]: [what was added] (source: [URL])
- ...

Proceed with these additions? (y) or edit:

Wait for confirmation before proceeding to Step 5. If no gaps found: skip silently, go straight to Step 5.

Step 4.7 — TikZ visualization

Read references/tikz-patterns.md for the reusable snippet library.

Scan restructured content for three visual triggers:

Source signalTikZ typePlacement
`mermaid blockFlowchart (pattern 1)Own frame
Equation with a curve/decay/relationship shapeCurve plot (pattern 2)Inline columns
Two-column or two-list comparisonTwo-box comparison (pattern 3)Inline columns
Timeline / sequential steps (>4)Timeline (pattern 4)Own frame

For own-frame diagrams only, show:

code
TikZ: Found [N] full-frame diagram(s):
- [description of each]
Add these slides? (y/n):

Wait for confirmation. Inline TikZ inserts silently.

Add to .tex preamble:

latex
\usepackage{tikz}
\usetikzlibrary{arrows.meta,shapes,positioning,calc}

Step 5 — Generate Beamer .tex

  • One \begin{frame} per logical section
  • Title frame with \titlepage
  • Use confirmed Beamer theme via \usetheme{}
  • Keep slide text concise (3–5 bullet points per frame max)
  • Output file: same directory as input, -presentation.tex suffix

Step 6 — Compile to PDF

powershell
# Windows (MiKTeX installed at):
$pdflatex = "C:\Users\User\AppData\Local\Programs\MiKTeX\miktex\bin\x64\pdflatex.exe"
& $pdflatex -interaction=nonstopmode "-output-directory=<dir>" <output>.tex
bash
# Linux/Mac:
pdflatex <output>.tex
xelatex <output>.tex

If LaTeX not found: deliver .tex with message:

"Compile with: pdflatex notes-presentation.tex or upload to Overleaf."

References

  • references/research.md — 22 frameworks with detect_signal keywords, decision tree, Beamer theme mappings

Install & Usage

1
Create the skills directory
mkdir -p .claude/skills
2
Download the skill file
mkdir -p .claude/skills && curl -o .claude/skills/present.md https://raw.githubusercontent.com/OrnlyP63/present-skill/main/SKILL.md
3
Invoke in Claude Code
/present
View source on GitHub

Frequently Asked Questions

What is present?

Transform raw .md files into Beamer LaTeX presentations (PDF). Auto-detects the best presentation framework (Minto Pyramid, SCR, STAR, BLUF, IMRaD, and 17 more) from content signals, restructures content into that framework's narrative flow, and generates a compilable Beamer .tex file. Use when user provides a .md file and asks to make a presentation, create slides, or export to PDF.

How to install present?

To install present, create the .claude/skills directory in your project, then run the curl command to download the skill file. Once installed, invoke it in Claude Code with /present.

What is present best for?

present is a community categorized under General. Created by OrnlyP63.