Lesson Generator Skill

Persona

Think like a curriculum designer who creates consistent, high-quality lessons that build progressively. You ensure every lesson has complete metadata, follows the 4-Layer Teaching Framework, and respects hardware tier constraints.

---

Pre-Flight Questions

Before generating any lesson content, ask yourself:

1. Context Gathering

• Q: What module and chapter is this lesson in?

  • Impact: Determines layer (L1-L4), proficiency, and hardware requirements

• Q: What lessons come before and after?

  • Impact: Ensures continuity and avoids redundancy

• Q: What is the target proficiency level?

  • A2 (Beginner): 5-7 concepts, heavy scaffolding
  • B1 (Intermediate): 7-10 concepts, moderate scaffolding
  • C2 (Advanced): No concept limits, minimal scaffolding

2. Metadata Completeness

• Q: Does this lesson have ALL required metadata fields?
  • See Required Metadata Fields below

3. Docusaurus Conventions

• Q: Am I following all Docusaurus conventions?
  • File extension: .md (not .mdx)
  • No Mermaid diagrams
  • No < in prose
  • Links use .md and README.md
  • See: docusaurus-conventions skill

---

Required Metadata Fields

Complete Frontmatter Template

---
id: lesson-{chapter}-{lesson}-{slug}
title: "Lesson {C}.{L}: {Title}"
sidebar_position: {N}
sidebar_label: "{C}.{L} {Short Title}"
description: "{One-line description for SEO and previews}"
duration_minutes: {45|60|75|90}
proficiency_level: "{A2|B1|C2}"
layer: "{L1|L2|L3|L4}"
hardware_tier: {1|2|3|4}
learning_objectives:
  - "{Objective 1 using Bloom's verb}"
  - "{Objective 2}"
  - "{Objective 3}"
skills:
  - "{skill-slug-1}"
  - "{skill-slug-2}"
cognitive_load:
  new_concepts: {N}
tier_1_path: "{Cloud fallback description}"
generated_by: "content-implementer v1.0.0"
created: "{YYYY-MM-DD}"
version: "1.0.0"
---

Field Descriptions

Field	Required	Description	Example
id	Yes	Unique across all docs	lesson-1-1-digital-to-physical
title	Yes	Full lesson title	"Lesson 1.1: From ChatGPT to Walking Robots"
sidebar_position	Yes	Order in sidebar	1
sidebar_label	Yes	Short label for sidebar	"1.1 Digital to Physical"
description	Yes	SEO/preview text	"Understanding differences between software and embodied AI"
duration_minutes	Yes	Expected completion time	45
proficiency_level	Yes	CEFR-aligned level	"A2", "B1", "C2"
layer	Yes	Teaching layer	"L1", "L2", "L3", "L4"
hardware_tier	Yes	Minimum hardware required	1 (cloud), 2 (GPU), 3 (Jetson), 4 (robot)
learning_objectives	Yes	Measurable outcomes (3-5)	Array of strings
skills	Yes	Related skill slugs	["ros2-fundamentals", "ros2-cli"]
tier_1_path	Recommended	Cloud fallback	"Cloud ROS 2 (TheConstruct)"
cognitive_load.new_concepts	Recommended	Concept count	7
generated_by	Yes	Generator attribution	"content-implementer v1.0.0"
created	Yes	Creation date	"2025-11-29"
version	Yes	Content version	"1.0.0"

Deprecated Fields (Do Not Use)

Deprecated	Use Instead
cefr_level	proficiency_level
duration	duration_minutes
estimated_time	duration_minutes
chapter / lesson	id with pattern

---

Principles

Principle 1: Metadata-First Generation

Always generate complete metadata BEFORE content.

The metadata defines:

• How the lesson appears in navigation
• What proficiency constraints apply
• What hardware is required
• What learning is measured

Principle 2: Layer-Appropriate Structure

L1 (Manual Foundation):

• Direct explanations with analogies
• Step-by-step walkthroughs
• NO AI collaboration yet
• Heavy scaffolding

L2 (AI Collaboration):

• Three Roles framework (invisible to student)
• Bidirectional learning examples
• "Try With AI" prompts
• Moderate scaffolding

L3 (Intelligence Design):

• Create reusable skills/patterns
• Encode tacit knowledge
• Cross-project value
• Minimal scaffolding

L4 (Spec-Driven):

• Specification-first approach
• Compose accumulated intelligence
• Capstone integration
• Professional autonomy

Principle 3: Hardware-Aware Content

Every lesson must specify:

1. Minimum hardware_tier (1-4)
2. tier_1_path fallback for cloud users
3. Hardware gates for tier-specific content

**Hardware Tier**: {N} | **Cloud Path**: {fallback description}

Principle 4: Consistent Naming

ID Pattern: lesson-{chapter}-{lesson}-{slug}

• lesson-1-1-digital-to-physical
• lesson-3-4-services-parameters
• lesson-7-3-testing-validation

Sidebar Label Pattern: "{C}.{L} {Short Title}"

• "1.1 Digital to Physical"
• "3.4 Services & Parameters"
• "7.3 Testing Validation"

---

Lesson Structure Template

---
[Complete frontmatter - see above]
---

# {Lesson Title}

**Duration**: {N} minutes | **Layer**: {L1-L4} | **Tier**: {N} ({description})

{Opening hook - 2-3 paragraphs motivating the topic}

## Learning Objectives

By the end of this lesson, you will be able to:
- {Objective 1}
- {Objective 2}
- {Objective 3}

## {Section 1: Core Content}

{Layer-appropriate content}

### {Subsection}

{Progressive explanation}

## {Section 2: Practice/Application}

{Hands-on content appropriate to layer}

## {Section 3: Assessment} (optional for L1)

{Understanding validation}

## Try With AI

{AI collaboration prompts - required for L2+}

---

**Previous:** [Link] | **Next:** [Link]

---

Checklist (Use Before Every Lesson)

Metadata

• All required fields present
• id is unique and follows pattern
• proficiency_level not cefr_level
• duration_minutes not duration
• skills array populated
• hardware_tier specified
• tier_1_path for tier > 1

Content

• Opening hook engages reader
• Learning objectives measurable (Bloom's verbs)
• Layer-appropriate teaching approach
• Code examples have outputs (for L2+)
• "Try With AI" section present (for L2+)

Docusaurus

• File extension .md
• No Mermaid diagrams
• No < in prose
• Links use .md extension
• Build passes

---

Integration

This skill is typically invoked by:

• content-implementer agent during /sp.implement
• Direct lesson creation workflows
• Module content generation

Dependencies:

• docusaurus-conventions - File naming and syntax
• learning-objectives - Bloom's taxonomy alignment
• assessment-builder - Proficiency-appropriate assessments

---

Version History

Version	Date	Change
0.1.0	2025-11-28	Initial placeholder
1.0.0	2025-11-29	Full implementation with metadata standards from Module 1