Presentation Master is not just a presentation generator—it's a presentation coach that teaches world-class design principles while creating slides. Built on decades of wisdom from Garr Reynolds, Nancy Duarte, Guy Kawasaki, Seth Godin, TED speakers, Edward Tufte, and Steve Jobs, this skill ensures every presentation meets the highest standards.

Key Innovation

Instead of just making slides, this skill:

Teaches presentation best practices through guided creation
Validates against proven design principles automatically
Adapts to different presentation types and audiences
Coaches you through checkpoints to ensure quality
Learns from each presentation to improve future recommendations

What Makes a Great Presentation?

Based on research of thousands of successful presentations, the masters agree on these principles:

Simplicity - Less is always more (Seth Godin: 6 words max per slide)
Visual Dominance - Show, don't write (Garr Reynolds: Pictures > text)
Story Structure - Random facts don't stick; stories do (Nancy Duarte: Sparkline)
One Idea Per Slide - Cognitive load is real (Guy Kawasaki: 10 slides max)
Audience as Hero - It's their transformation, not your information (Duarte)
Data Integrity - Visual truth matters (Edward Tufte: Lie factor 0.95-1.05)

Features

Guided Creation with Checkpoints

Phase 1: Discovery & Context Analysis

Analyzes topic, audience, duration, purpose
Detects presentation type (board update, keynote, training, TED-style)
Recommends story framework

Phase 2: Content Development

Researches topic if needed
Extracts key concepts (max 10 per Kawasaki)
Drafts slide outline with one concept per slide
Identifies slides needing visuals

Phase 3: Visual Strategy

Proposes image types (infographic, photo, diagram, data viz)
Suggests visual concepts and metaphors
Estimates generation costs
Gets approval before generating

Phase 4: Generation & Validation

Generates approved images in parallel
Builds slides using selected format (PPTX/Slides/Canva)
Applies design rules automatically
Runs validation scoring (0-100)
Generates quality report

Phase 5: Iteration & Refinement (if needed)

Applies requested changes
Re-validates quality
Saves to Obsidian with metadata
Documents learnings

Automatic Quality Validation

Every presentation is scored 0-100 against these criteria:

Simplicity (10pts): Word count, visual clutter
Visual Dominance (10pts): Image quality, text-to-visual ratio
Story Structure (10pts): Narrative arc, emotional beats
One Idea/Slide (10pts): Concept clarity
Typography (8pts): Size (30pt+ minimum), consistency
Layout (7pts): Hierarchy, whitespace, alignment
Color/Contrast (7pts): Readability (4.5:1 minimum)
Media Quality (8pts): Image resolution, relevance
Cognitive Load (20pts): Mayer's 12 multimedia principles
Data Integrity (10pts): Tufte principles (if applicable)

Target Score: 85+ for high-quality presentations

Critical Violations (Auto-Fail)

Font size < 30pt
10 core concepts
Bullet points detected
Paragraphs (>2 sentences)
Poor contrast (<4.5:1)
Default templates

Multi-Format Support

Phase 1 (Current): PPTX

Uses existing pptx skill
Direct PptxGenJS API for simple slides
Saves to Obsidian Research folder

Phase 2 (Coming): Google Slides

Google Slides API v1
Supports all 3 accounts (psd, kh, hrg)
Returns shareable link

Phase 3 (Future): Canva

Canva REST API
Auto-applies brand kits

Slide Patterns

Six proven patterns that form the foundation of world-class presentations:

1. Title Slide

Purpose: Set tone, introduce topic Elements: 3-6 words, strong visual or clean typography Best for: Opening, section breaks

2. One Big Idea

Purpose: Maximum impact, single concept Elements: 1-3 words max, 60-120pt font Best for: Key insights, memorable moments (Seth Godin's 6-word rule)

3. Visual + Caption

Purpose: Visual storytelling with context Elements: Large high-quality image (70-80%), short caption (1 line) Best for: Concepts, examples, emotional moments

4. Data Visualization

Purpose: Show quantitative information clearly Elements: Clean charts following Tufte principles Best for: Evidence, trends, comparisons

5. Timeline / Process

Purpose: Show progression, sequence, chronology Elements: 3-7 steps, visual connections, minimal text Best for: Processes, roadmaps, historical progression

6. Transition / Section Break

Purpose: Signal major shifts Elements: 1-5 words, distinct visual treatment Best for: Moving between major sections

See principles/slide-patterns.md for comprehensive details on each pattern.

Story Frameworks

Choose the right narrative structure for your presentation:

Nancy Duarte's Sparkline

Duration: 20-30 minutes Slides: 18-25 Best for: Keynotes, vision casting, inspirational talks Structure: Alternate "what is" (reality) with "what could be" (aspiration)

Steve Jobs' Rule of Three

Duration: 15-30 minutes Slides: 12-18 Best for: Product launches, demos, persuasive presentations Structure: Three main sections, three points per section

TED Talk Structure

Duration: 15-20 minutes Slides: 12-18 Best for: Idea-centric, inspirational talks Structure: Hook → Personal → Core Idea → Call to Action → Close

Classic Three-Act Structure

Duration: 20-45 minutes Slides: 15-35 Best for: Case studies, transformation narratives Structure: Setup (25%) → Confrontation (50%) → Resolution (25%)

See principles/story-frameworks.md for detailed breakdowns.

Adaptive Pattern Selection

The skill automatically recommends pattern distribution based on presentation type:

Board Update: 40% Data Viz, 30% Visual+Caption, 20% Transitions, 10% Titles Keynote: 40% Big Ideas, 30% Visual+Caption, 20% Titles, 10% Transitions Training: 40% Process, 30% Data Viz, 20% Visual+Caption, 10% Titles Pitch: 30% Data Viz, 30% Big Ideas, 25% Visual+Caption, 15% Titles/Transitions

Image Generation Integration

Intelligent visual recommendations based on slide content:

Numeric data → Infographic
Abstract concepts → Conceptual image/metaphor
Concrete examples → Realistic photography
Processes/flows → Diagrams
Trends/comparisons → Data visualization

Cost Management:

Shows estimated cost before generating ($0.13-$0.24 per image at 2K)
Allows opt-out of specific images
Tracks spending in presentation metadata

Naming Convention:

{presentation-id}-slide-{number}-{type}.png
Example: cybersecurity-2025-slide-02-threat-timeline.png

Brand Integration

Automatically applies branding:

PSD presentations: Uses psd-brand-guidelines skill
Personal presentations: Uses personal brand (if defined)
Client presentations: Custom brand kit (future)

File Structure

presentation-master/
├── SKILL.md                           # Main skill documentation
├── README.md                          # This file
├── package.json                       # Dependencies
├── principles/
│   ├── masters.md                     # Best practices from experts (1,200+ lines)
│   ├── validation-rules.md            # Quality scoring system
│   ├── slide-patterns.md              # Template library (3,500+ lines)
│   └── story-frameworks.md            # Narrative structures (5,000+ lines)
├── scripts/
│   ├── analyze-context.js             # ✅ Detect presentation type & recommend framework
│   ├── validate-presentation.js       # ✅ Score presentation 0-100
│   └── suggest-images.js              # ✅ Visual recommendations & cost estimation
└── adapters/
    ├── pptx-adapter.js                # ✅ PowerPoint translation layer
    ├── slides-adapter.js              # ✅ Google Slides translation layer
    └── canva-adapter.js               # (future) Canva integration

Usage Examples

Example 1: Board Update

User: "Create presentation on cybersecurity trends for school board, 15 minutes"

Geoffrey:
**Context Analysis**
- Type: Board/Executive Update
- Audience: School board (non-technical)
- Duration: 15 minutes
- Purpose: Inform + Recommend action

**Recommended Structure**
Current State → Trends → Implications → Recommendations (10 slides)

[After approval]

**Draft Outline**
1. Title: Cybersecurity Trends 2025
2. Current Threat Landscape [visual: threat timeline]
3. Trend 1: AI-Powered Attacks [visual: concept illustration]
...

**Visual Strategy**
- 7 images recommended
- Estimated cost: $0.94 (2K resolution)
- Mix: 3 infographics, 2 charts, 1 diagram, 1 illustration

[After approval and generation]

**Quality Score: 91/100**
✓ All fonts 36pt+
✓ Avg 4.2 words/slide
✓ High contrast (8.5:1)
⚠ Slide 8 has 7 words (rec: 6 max)

Created: /Obsidian/Research/cybersecurity-trends-2025-board.pptx

Example 2: Keynote

User: "Create keynote on future of education for conference, 30 minutes"

Geoffrey:
**Context Analysis**
- Type: Keynote/TED-style
- Audience: Education professionals + parents
- Duration: 30 minutes
- Purpose: Inspire + Shift perspective

**Recommended Structures**
1. Sparkline: What is → What could be (12 slides) ⭐
2. Rule of Three: Past → Present → Future (9 slides)
3. Three-Act: Problem → Trials → New reality (10 slides)

[Continues with guided creation...]

Dependencies

pptxgenjs: PowerPoint generation library (current Phase 1)
image-gen skill: Visual content generation
psd-brand-guidelines skill: Brand colors/logos
research skill: Content development
obsidian-manager skill: Knowledge storage

Installation

This skill is part of Geoffrey's skill library. No separate installation needed.

Ensure dependencies are installed:

cd /Users/hagelk/non-ic-code/geoffrey/skills/presentation-master
bun install

Triggers

Natural language activation:

"create presentation"
"make presentation"
"build presentation"
"design deck"
"make slides"
"create slides"
"presentation for..."
"slides for..."

Version

Current: 1.2.0 (Phases 1-3 Complete)

✅ Phase 1: PPTX support via pptx skill
✅ Phase 2: Visual Intelligence (analyze-context, suggest-images, validate-presentation scripts)
✅ Phase 3: Google Slides support via google-workspace skill Future: 1.3.0 (Canva integration)

Learning & Iteration

After each presentation, metadata is saved to Obsidian:

{
  "presentation_id": "cybersecurity-2025-board",
  "type": "board-update",
  "duration": "15min",
  "slides": 10,
  "quality_score": 91,
  "images_generated": 7,
  "generation_cost": 0.94,
  "user_edits": ["slide-8-wording"],
  "effectiveness": "approved-minor-changes",
  "created": "2025-11-24",
  "learnings": "Board prefers data charts over concept graphics"
}

This enables:

Quality score trending
Pattern library expansion
User preference learning
Cost optimization

Philosophy

Code Before Prompts (Geoffrey Principle):

Deterministic design rules codified in software
AI orchestrates, doesn't improvise design
Validation is automatic, not subjective

Progressive Disclosure (Geoffrey Principle):

Tier 1: Core workflow (always loaded)
Tier 2: Detailed principles (loaded when skill activates)
Tier 3: Scripts and data (fetched just-in-time)

Checkpoints Are Non-Negotiable:

Structure approval before content development
Outline approval before visual strategy
Visual approval before generation
Quality review before finalization

References

Books:

Presentation Zen by Garr Reynolds
Resonate by Nancy Duarte
slide:ology by Nancy Duarte
The Visual Display of Quantitative Information by Edward Tufte
Multimedia Learning by Richard Mayer

Articles:

Guy Kawasaki: The 10/20/30 Rule of PowerPoint
Seth Godin: Really Bad PowerPoint
TED: Create + Prepare Slides Guide

Principles Documentation:

principles/masters.md - Comprehensive best practices
principles/validation-rules.md - Quality scoring details
principles/slide-patterns.md - Pattern specifications
principles/story-frameworks.md - Narrative structures

Future Enhancements

Phase 4: Canva Integration

Canva REST API adapter
Auto-apply brand kits
Template library integration

Phase 5: Advanced Features

Auto-generate speaker notes
Rehearsal timing suggestions
A/B test different structures
Effectiveness tracking
Personal pattern library from successes

Contributing

Improvements welcome! When adding new features:

Follow existing architectural patterns
Document in appropriate principles file
Add examples to SKILL.md
Update version in package.json
Add learning to metadata tracking

License

MIT License - See LICENSE file for details

Remember: The best presentations are simple, visual, story-driven, and focused on the audience's transformation—not the speaker's information.

Built with ❤️ by Geoffrey