Course Creator Skills

Blockchain Reference Implementation

The Blockchain course is the flagship implementation of the course-creator system. It demonstrates the full pipeline output including all 7 stages.

Quick Start — Minimal Course

A minimal 3-lecture course.yaml to get started. This covers one module pair and one classification lecture, a single project, and basic quiz settings.

course_name: "Introduction to Machine Learning"
short_name: "IntroML"
academic_year: "2025/2026"
semester: "Winter"
institution: "Example University"
instructor:
  name: "Dr. Jane Smith"
  email: "jane.smith@example.edu"
website_url: "https://your-org.github.io/IntroML/"
github_repo: "IntroML"

modules:
  - name: "Foundations"
    lectures:
      - title: "What is Machine Learning?"
        tex_file: "lecture-01-intro.tex"
        topics: ["supervised learning", "unsupervised learning", "history"]
      - title: "Linear Regression"
        tex_file: "lecture-02-regression.tex"
        topics: ["regression", "gradient descent", "loss functions"]
  - name: "Classification"
    lectures:
      - title: "Logistic Regression & SVMs"
        tex_file: "lecture-03-classification.tex"
        topics: ["classification", "SVM", "decision boundary"]

projects:
  - title: "Predict Housing Prices"
    description: "Apply regression techniques to the Boston housing dataset"
    lectures: [1, 2]
    difficulty: "beginner"

quizzes:
  per_lecture: 5
  advanced_ratio: 0.3

Step-by-Step Generation

How to generate a course from start to deployed GitHub Pages site.

# Step 1: Create your course directory
mkdir IntroML && cd IntroML

# Step 2: Create your course.yaml (as above)

# Step 3: Write your first lecture .tex files
# (Use the beamer-slide-creator skill in Claude Code)

# Step 4: Run the full pipeline
# In Claude Code, say:
# "Generate the complete IntroML course from course.yaml"
# The course-creator orchestrator handles all 7 stages automatically.

# Step 5: Preview locally
# Open the generated index.html in your browser

# Step 6: Deploy
# The deploy stage creates a GitHub repo and enables Pages

Running Individual Stages

You can run any single stage independently when iterating on specific content.

# Run only slides generation
"Generate slides for IntroML lecture 1 using course.yaml"

# Run only charts
"Generate charts for IntroML lecture 2"

# Run only galleries (after slides exist)
"Create slide galleries for all IntroML lectures"

# Run only deployment (after all content exists)
"Deploy the IntroML course to GitHub Pages"

Customization Guide

Override Chart Types

Specify chart types per lecture to override the auto-derived defaults.

modules:
  - name: "Finance"
    lectures:
      - title: "Stock Markets"
        tex_file: "lecture-01.tex"
        topics: ["stock prices", "trading volume"]
        charts:
          - type: "candlestick"
            title: "AAPL Price History"
          - type: "line_plot"
            title: "Trading Volume Over Time"

Custom Quiz Distribution

Increase quiz count and adjust the Bloom's Taxonomy level distribution.

quizzes:
  per_lecture: 8
  advanced_ratio: 0.5
  bloom_levels: ["analyze", "evaluate", "create"]

Custom Theme Colors

Override the default color palette applied to the generated site.

theme:
  primary_color: "#1a365d"
  secondary_color: "#2d3748"
  font_family: "Roboto"

Organization Repos Using the System

Tips & Best Practices

• Start with 2–3 lectures to test the pipeline before scaling up to a full 12-lecture course.
• Use the --dry-run flag in the deploy stage to preview changes before pushing to GitHub.
• Keep .tex files in a consistent directory structure; the orchestrator resolves paths relative to course.yaml.
• Run individual stages when iterating on specific content — you do not need to regenerate the whole course every time.
• The orchestrator handles stage dependencies automatically; simply run the full pipeline and let it sequence the work.