AI Lesson Planner

Chat-first toolkit for planning complete courses and generating lesson artifacts with specialized Copilot agents.

Recommended start (chat-first)

Use @course-planner as the first step to design the whole course, then run lesson-level agents.

@course-planner Plan a course on Python programming for upper-secondary students, 12 lessons, 90 minutes each.

What @course-planner prepares:

• course outline in lessons/README.md
• course metadata in .lesson-config.json
• lesson scaffolding checklist with ./scripts/generate-lesson.sh
• optional full-course scaffolding with ./scripts/generate-course-scaffolding.sh
• handoff prompts for lesson-level agents

Agent flow

1. @course-planner → design course structure (outline only)
2. @discussion-moderator → generate lesson highlights
3. @lesson-planner (optional) → expand highlights into discourse
4. @slides-maker → generate MARP slide files
5. ./scripts/generate-slides-pdf.sh → export final PDF

Quick start

1) Plan the course in chat

@course-planner Create a 10-lesson course on AI agents. Include prerequisites and progression by module.

2a) Scaffold lesson folders

Run generated commands (example):

./scripts/generate-lesson.sh 01 intro-ai-agents "Introduction to AI Agents"
./scripts/generate-lesson.sh 02 prompting-basics "Prompting Basics"

Or scaffold the whole course directly from .lesson-config.json:

./scripts/generate-course-scaffolding.sh

2b) Generate lesson content in chat

@discussion-moderator Prepare lesson #01: Introduction to AI Agents
@lesson-planner Create discourse for lesson #01
@slides-maker Create slides for lesson #01

3) Export PDF (per single lesson)

./scripts/generate-slides-pdf.sh lessons/lesson-01-intro-ai-agents

Agents

• course-planner: course outline, metadata, scaffolding plan (no lesson content)
• discussion-moderator: structured lesson highlights via moderated debate
• lesson-planner: detailed discourse from highlights
• slides-maker: MARP slide generation
• co-teacher: perspective agent used by moderator flow
• repository-creator (optional): publishing support

How does it work

Language behavior

• Default language is English.
• If the user explicitly requests another language, agents should follow that request.
• For course-level planning, language can be derived from .lesson-config.json metadata when available.

Scripts

See full scripts documentation in scripts/README.md.

License

This project is licensed under the GNU General Public License v3.0 (or later). See LICENSE for full terms.