learning-content-creator

name	learning-content-creator
description	Transform research materials into structured learning content with bilingual support. Use when users want to create learning content from research, convert research to learning paths, make learning materials, or translate learning content between languages.

Learning Content Creator

Transform multi-model research materials into structured learning content, with bilingual support (English + Korean).

Quick Start

Workflow Overview

Research Materials → Learning Path (EN) → Translation (KO) → Frontmatter

Phase	Input	Output
1. Analyze	research/*.md files	Content outline
2. Create EN	Synthesized insights	learning/*.en.md
3. Translate KO	English content	learning/*.ko.md
4. Frontmatter	All files	YAML metadata added

Workflow Routing

Intent	Workflow
Create learning content from research	workflows/create-learning.md
Translate EN to KO	workflows/translate.md

Note: Phase 4 (Frontmatter) uses the doc-frontmatter schema. The calling command coordinates the multi-skill pipeline.

Phase 1: ANALYZE Research

Goal: Understand and synthesize multi-model research.

Input Structure

docs/NN-topic/
├── research/
│   ├── 00-research-prompt.en.md   # Original prompt
│   ├── 01-claude.en.md            # Claude's response
│   ├── 02-gpt/                    # GPT's response (may be multi-file)
│   │   ├── 01-concepts.en.md
│   │   ├── 02-relationships.en.md
│   │   └── ...
│   └── 03-gemini.en.md            # Gemini's response

Analysis Checklist

Check	Question
Coverage	What topics do all models agree on?
Unique	What unique insights does each model provide?
Conflicts	Where do models disagree? How to resolve?
Gaps	What's missing? What questions remain?

Output: Content Outline

# Learning Content Outline

## Modules (6-8 recommended)

| # | Module | Topics | Sources |
|---|--------|--------|---------|
| 1 | Fundamentals | Definitions, core concepts | All models |
| 2 | Relationships | How parts connect | GPT, Claude |
| ... | ... | ... | ... |

## Key Insights by Source

- **Claude**: [unique insights]
- **GPT**: [unique insights]
- **Gemini**: [unique insights]

## Synthesis Strategy

[How to combine insights without redundancy]

Phase 2: CREATE English Content

Goal: Create structured learning modules in English.

Directory Structure

docs/NN-topic/
└── learning/
    ├── README.en.md           # Course overview, learning path
    ├── 01-module-name.en.md   # Module 1
    ├── 02-module-name.en.md   # Module 2
    └── ...

Module Template

# Module N: Title

> One-sentence module summary

## Learning Objectives

After completing this module, you will:
- [Objective 1]
- [Objective 2]
- [Objective 3]

---

## N.1 First Section

[Content]

## N.2 Second Section

[Content]

---

## Key Takeaways

- [Takeaway 1]
- [Takeaway 2]
- [Takeaway 3]

## Exercises

### Exercise N.1: [Name]

[Exercise description]

---

## Next Steps

Continue to [Module N+1: Title](./0N+1-title.en.md)

README Template

# Course Title

> Course tagline (one sentence)

## Course Overview

[2-3 paragraph description]

## Who This Is For

- [Audience 1]
- [Audience 2]

## Prerequisites

- [Prerequisite 1]
- [Prerequisite 2]

---

## Course Modules

| # | Module | Duration | Description |
|---|--------|----------|-------------|
| 1 | [Title](./01-name.en.md) | NN min | Description |
| 2 | [Title](./02-name.en.md) | NN min | Description |

**Total Time:** ~N hours

---

## Learning Path

### Beginner Track
[Visual flow diagram]

### Advanced Track
[Visual flow diagram]

---

## Source Materials

| Source | Description |
|--------|-------------|
| [Research Prompt](../research/00-research-prompt.en.md) | Original prompt |
| [Claude](../research/01-claude.en.md) | Claude's response |
| [GPT](../research/02-gpt/) | GPT's response |
| [Gemini](../research/03-gemini.en.md) | Gemini's response |

Writing Guidelines

Guideline	Description
Synthesize	Don't copy verbatim; synthesize insights
Attribute	Note which model contributed which insight
Practical	Focus on actionable knowledge
Consistent	Use same terminology throughout
Progressive	Build complexity gradually

Phase 3: TRANSLATE to Korean

Goal: Create high-quality Korean translations.

Naming Convention

English	Korean
*.en.md	*.ko.md
README.en.md	README.ko.md
01-fundamentals.en.md	01-fundamentals.ko.md

Translation Guidelines

Aspect	Guideline
Technical terms	Keep English for universally used terms (e.g., Command, Skill, Agent)
Headers	Translate headers
Code blocks	Keep code in English, translate comments
Tables	Translate content, keep structure
Links	Update to point to .ko.md counterparts

Do NOT Translate

• Code snippets
• File paths
• Command examples
• Technical identifiers (kebab-case names, etc.)

Translation Checklist

• All .en.md files have .ko.md counterparts
• Internal links updated to .ko.md versions
• Technical terms consistently handled
• Tables and diagrams preserved
• Exercises and examples localized where appropriate

Phase 4: ADD Frontmatter

Goal: Add YAML frontmatter to all learning documents.

Frontmatter Schema

Use the frontmatter schema from doc-frontmatter skill (see .claude/skills/doc-frontmatter/references/schema.md).

Frontmatter Template for Learning Content

---
title: "Module Title"
description: "50-160 char summary of what this module teaches"
type: tutorial
tags: [AI, Architecture, BestPractice]
order: 1
depends_on: [./prerequisite-module.en.md]
related: [./related-module.en.md]
---

Type Selection for Learning

Content Type	type Value
Course overview (README)	index
Step-by-step module	tutorial
Reference/spec	reference
Concept explanation	explanation

Execution

For each file:

1. Extract title from H1
2. Generate description from first paragraph
3. Determine type based on content
4. Select relevant tags (max 5)
5. Set order from filename prefix
6. Add depends_on/related if applicable

Quality Checklist

Phase 1: Analyze

• All research files read
• Key insights extracted from each model
• Conflicts identified and resolved
• Content outline created

Phase 2: Create EN

• All modules follow template
• Learning objectives are measurable
• Content synthesizes (not copies) sources
• Exercises included in each module
• Links work correctly

Phase 3: Translate KO

• All files translated
• Technical terms consistent
• Links updated to .ko.md
• Natural Korean (not machine-translation quality)

Phase 4: Frontmatter

• All files have frontmatter
• Required fields present (title, description, type)
• Tags from controlled vocabulary
• Dependencies correctly specified

Example Output

See docs/01-structure-organizer/learning/ for a complete example:

docs/01-structure-organizer/learning/
├── README.en.md              # Course index (EN)
├── README.ko.md              # Course index (KO)
├── 01-fundamentals.en.md     # Module 1 (EN)
├── 01-fundamentals.ko.md     # Module 1 (KO)
├── 02-relationships.en.md    # Module 2 (EN)
├── 02-relationships.ko.md    # Module 2 (KO)
├── 03-decision-framework.en.md
├── 03-decision-framework.ko.md
├── 04-templates.en.md
├── 04-templates.ko.md
├── 05-examples.en.md
├── 05-examples.ko.md
├── 06-anti-patterns.en.md
└── 06-anti-patterns.ko.md