Step 1: Syllabus Creation
Goal
Create a structured outline of all topics that will be covered in the guide. This becomes the backbone of the entire document.
Inputs Required
| Input | Location | Purpose |
|-------|----------|---------|
| Main articles | context_engineering_new/1_*.md through 9_*.md | Core content |
| LinkedIn posts | context_engineering_new/linkedin_post_*.md | Key highlights |
| Interview Q&A | context_engineering/_chatgpt/otazky.md | Raw insights |
| Article plan | context_engineering/_chatgpt/clanky.md | Original structure |
| Presentation | codecon_materials/presentation_structure.md | Condensed version |
Process
1.1 Content Inventory
First, create an inventory of all existing content:
## Content Inventory
### Articles
| # | File | Title | Main Topics | Word Count |
|---|------|-------|-------------|------------|
| 1 | 1_prompt_is_not_enough.md | ... | ... | ... |
| 2 | ... | ... | ... | ... |
### Supporting Materials
| File | Type | Key Content |
|------|------|-------------|
| otazky.md | Interview | 25+ Q&A pairs |
| ... | ... | ... |
Action: Read each file, extract title and list 3-5 main topics covered.
1.2 Topic Extraction
From the inventory, extract all unique topics mentioned:
## Raw Topic List
1. Prompt vs Context (definition)
2. Mental model of AI as colleague
3. Common mistakes with AI
4. Atomization / breaking down tasks
5. Iteration and feedback
6. Context contamination
7. Tools comparison
8. Team implementation
9. Vibe Coding critique
10. ...
Action: Don't filter yet. List everything. Duplicates are okay at this stage.
1.3 Topic Clustering
Group related topics into logical clusters:
## Topic Clusters
### Cluster A: Foundations
- What is context engineering
- Prompt vs Context difference
- Why context matters
- Mental model (AI as colleague)
### Cluster B: How to Do It
- Breaking tasks into atomic parts
- Providing good context
- Examples and templates
- Iteration and feedback
### Cluster C: Common Problems
- Mistakes people make
- Context contamination
- When to start fresh
### Cluster D: Tools & Practice
- Tool recommendations
- Workflows
- Real examples
### Cluster E: Bigger Picture
- Team implementation
- Career implications
- Industry perspective (Vibe Coding)
Action: Each cluster will potentially become a Part or Section of the guide.
1.4 Hierarchy Definition
Convert clusters into a hierarchical structure:
## Syllabus Draft v1
# Part 1: Understanding Context Engineering
## Chapter 1: Why Prompt Is Not Enough
### 1.1 The difference between prompt and context
### 1.2 Mental model: AI as a new colleague
### 1.3 Why most people fail with AI
## Chapter 2: What Good Context Looks Like
### 2.1 Components of good context
### 2.2 The task template
### 2.3 Examples of bad vs good context
# Part 2: The Practice
## Chapter 3: Breaking Down Problems
### 3.1 Atomization principle
### 3.2 The 1-2 minute rule
### 3.3 When tasks are too big
... (continue for all chapters)
Action: Aim for 3-5 Parts, 2-4 Chapters per Part, 2-4 Sections per Chapter.
1.5 Coverage Mapping
Map source content to syllabus topics:
## Coverage Map
| Syllabus Topic | Primary Source | Supporting Sources |
|----------------|----------------|-------------------|
| 1.1 Prompt vs Context | Article 1, Interview Q1-3 | Presentation slide 1.2 |
| 1.2 AI as colleague | Article 1, Interview Q8 | LinkedIn post 1 |
| 2.1 Components | Article 2 | Interview Q4-7 |
| ... | ... | ... |
Action: Every topic must have at least one primary source. If a topic has no source, either remove it or mark it as "needs original content."
Output Format
Create file: syllabus.md
# Context Engineering Guide - Syllabus
## Version: 1.0
## Date: [date]
## Status: Draft
---
## Structure
# Part 1: [Part Title]
## Chapter 1: [Chapter Title]
### 1.1 [Section Title]
- Source: [file reference]
- Key points: [bullet list]
### 1.2 [Section Title]
- Source: [file reference]
- Key points: [bullet list]
## Chapter 2: [Chapter Title]
...
# Part 2: [Part Title]
...
---
## Coverage Summary
| Part | Chapters | Sections | Sources Used |
|------|----------|----------|--------------|
| 1 | 2 | 6 | Articles 1-2, Interview |
| 2 | 3 | 9 | Articles 3-5 |
| ... | ... | ... | ... |
---
## Unmapped Content
Content from sources that doesn't fit current structure:
- [Item 1] from [source] - consider for: [suggestion]
- [Item 2] from [source] - consider for: [suggestion]
---
## Open Questions
1. Should [topic X] be separate chapter or merged with [topic Y]?
2. Is [topic Z] too detailed for this guide?
Quality Criteria
| Criterion | Check | |-----------|-------| | All source files reviewed | ☐ | | No major topic missing | ☐ | | Logical flow from basic to advanced | ☐ | | Each chapter has clear purpose | ☐ | | No orphan topics (uncategorized) | ☐ | | Coverage map complete | ☐ | | Estimated length reasonable (50-80 pages) | ☐ |
Common Pitfalls
- Too granular - Don't make every paragraph a section. Keep it manageable.
- Too broad - Chapters should be specific enough to have clear scope.
- Missing the thread - The syllabus should tell a story: problem → understanding → solution → practice.
- Ignoring source material - If something is in the articles but not in syllabus, you're losing content.
Example: Good vs Bad Syllabus
Bad:
Chapter 1: AI
Chapter 2: Context
Chapter 3: Tools
Chapter 4: Examples
Problem: Too vague, no story, unclear what each chapter covers.
Good:
Part 1: The Foundation
Chapter 1: Why Your AI Prompts Fail (And What to Do Instead)
1.1 The "giving little, expecting much" trap
1.2 Understanding what AI actually needs
1.3 The mental shift from prompting to context engineering
Chapter 2: The Anatomy of Good Context
2.1 The five components every context needs
2.2 The universal task template
2.3 Before/after: Real examples that work
Better: Clear progression, specific titles, reader knows what to expect.
Time Estimate
- Content inventory: 30-45 minutes
- Topic extraction: 20-30 minutes
- Clustering: 20-30 minutes
- Hierarchy creation: 30-45 minutes
- Coverage mapping: 30-45 minutes
Total: 2-3 hours
Next Step
→ Once syllabus.md is complete, proceed to 02_topic_analysis.md