/**
* @fileoverview Auto-generated template definitions
* DO NOT EDIT MANUALLY - Generated by scripts/embed-templates.ts
* @module src/notebook/templates.generated
*/
export const TEMPLATES = {
'sequential-feynman': `<!-- srcbook:{"language":"[LANGUAGE]"} -->
# π§ Sequential Feynman Learning: [TOPIC]
**Topic**: \`[TOPIC]\`
**Started**: [DATE]
**Current Phase**: Phase 1 - Research
**Refinement Cycle**: N/A
---
## Progress Checklist
- [ ] Phase 1: Research & Synthesis completed
- [ ] Phase 2: Initial Feynman Explanation written
- [ ] Phase 3, Cycle 1: Major issues identified & fixed
- [ ] Phase 3, Cycle 2: Depth added & polished
- [ ] Phase 3, Cycle 3: Final validation passed
- [ ] Phase 4: Expert skill created
---
# Phase 1: Research & Synthesis
## What I Need to Learn
*Before diving into sources, clarify what I need to understand:*
- **Core Purpose**: Why does this technology/concept exist?
- **Key Mechanisms**: How does it actually work?
- **Mental Models**: What are the governing principles?
- **Usage Patterns**: How do practitioners actually use this?
- **Anti-Patterns**: What are common mistakes?
- **Trade-offs**: What are the costs and benefits?
---
## Sources & Key Findings
*Track authoritative sources and main takeaways:*
### Source 1: [Title/URL]
- Key point 1
- Key point 2
- Questions raised:
### Source 2: [Title/URL]
- Key point 1
- Key point 2
- Questions raised:
### Source 3: [Title/URL]
- Key point 1
- Key point 2
- Questions raised:
*Add more sources as needed*
---
## π€ What I Don't Understand Yet
*Be explicit about gaps - they're where the learning happens:*
1.
2.
3.
*This list should shrink as you progress through the notebook*
---
# Phase 2: Feynman Explanation
## What We're Learning Today
*Start with a relatable scenario that hooks interest:*
[Imagine you're explaining this to an intelligent college freshman who asks "Why should I care about this?" - answer that question in 2-3 paragraphs using a concrete scenario they'd relate to]
**Example structure:**
- Relatable problem or scenario
- Why existing approaches fall short
- How this concept solves it
---
## The Core Mental Model: [Your Analogy]
*Pick ONE central analogy that captures the essence - restaurants, mail systems, orchestras, etc.*
**The Analogy:**
[Develop your analogy in detail - what maps to what?]
**The Mapping:**
| Real System Component | Analogy Equivalent | Why This Maps |
|----------------------|-------------------|---------------|
| | | |
| | | |
**Limitations of This Analogy:**
*Be honest about where the analogy breaks down:*
-
---
## Concept 1: [Concept Name]
**In simple terms:**
[Explain using the analogy or simple language]
**Why this matters:**
[Connect to the larger picture]
**Concrete example:**
[Show a real, specific case]
---
## Concept 2: [Concept Name]
**In simple terms:**
[Explain using the analogy or simple language]
**Why this matters:**
[Connect to the larger picture]
**Concrete example:**
[Show a real, specific case]
---
## Concept 3: [Concept Name]
**In simple terms:**
[Explain using the analogy or simple language]
**Why this matters:**
[Connect to the larger picture]
**Concrete example:**
[Show a real, specific case]
---
## Concept 4: [Concept Name]
**In simple terms:**
[Explain using the analogy or simple language]
**Why this matters:**
[Connect to the larger picture]
**Concrete example:**
[Show a real, specific case]
---
## Concept 5: [Concept Name]
**In simple terms:**
[Explain using the analogy or simple language]
**Why this matters:**
[Connect to the larger picture]
**Concrete example:**
[Show a real, specific case]
---
## Concept 6: [Concept Name]
**In simple terms:**
[Explain using the analogy or simple language]
**Why this matters:**
[Connect to the larger picture]
**Concrete example:**
[Show a real, specific case]
---
\`\`\`typescript
// If applicable: Concrete code example demonstrating the concept
// Example:
// [Show actual working code that demonstrates what you just explained]
// Expected output:
// [What should happen when this runs?]
\`\`\`
---
## Summary: What We Learned
*Recap the key insights without jargon:*
1. **Main Idea**:
2. **Why It Matters**:
3. **How It Works**:
4. **When to Use It**:
5. **When NOT to Use It**:
**The One-Sentence Explanation:**
[If you had to explain this entire topic in one sentence, what would it be?]
---
# Phase 3: Refinement Through Thoughtbox Analysis
*This section tracks three refinement cycles. Each cycle has:*
- Pre-analysis questions to focus attention
- Space for thoughtbox analysis results
- Action items for refinement
- Post-refinement validation
---
## π Cycle 1: Finding Major Issues
**Current Status**: [Not Started | In Progress | Complete]
**Focus Areas for This Cycle:**
- Technical accuracy errors
- Major conceptual omissions
- Unclear or confusing explanations
- Missing critical examples
- Broken or weak analogies
**Prompt for thoughtbox:**
> "Analyzing this Feynman explanation of [TOPIC] to identify technical errors, major conceptual gaps, unclear sections, and missing critical examples. Will be thorough and critical."
---
### Issues Identified
*After running thoughtbox, document what needs fixing:*
#### Technical Errors
1.
2.
#### Conceptual Gaps
1.
2.
#### Unclear Explanations
1.
2.
#### Missing Examples
1.
2.
---
### Action Items
- [ ] Fix issue 1
- [ ] Fix issue 2
- [ ] Fix issue 3
---
### Refinements Made
*Document what you changed and why:*
1. **Fixed**: [What changed]
- **Why**: [Reasoning]
2. **Added**: [What was added]
- **Why**: [Reasoning]
---
*Now go back and update the Feynman Explanation section above β*
---
## π― Cycle 2: Adding Depth & Polish
**Current Status**: [Not Started | In Progress | Complete]
**Focus Areas for This Cycle:**
- Flow and organization
- Clarity of technical concepts
- Completeness of coverage
- Appropriateness of language for audience
- Strength and consistency of analogies
- Integration between sections
**Prompt for thoughtbox:**
> "Reviewing refined notebook for [TOPIC]. Will assess flow, clarity of technical concepts, completeness of coverage, and whether language is appropriate for an intelligent college freshman."
---
### Depth Opportunities
*What could be richer, clearer, or more complete?*
#### Flow Issues
1.
2.
#### Clarity Improvements
1.
2.
#### Completeness Gaps
1.
2.
#### Language Adjustments
1.
2.
---
### Action Items
- [ ] Improve area 1
- [ ] Improve area 2
- [ ] Improve area 3
---
### Refinements Made
*Document depth improvements:*
1. **Enhanced**: [What improved]
- **How**: [Specific changes]
2. **Integrated**: [What connected better]
- **How**: [Specific changes]
---
*Update the Feynman Explanation section above β*
---
## β¨ Cycle 3: Final Validation
**Current Status**: [Not Started | In Progress | Complete]
**Focus Areas for This Cycle:**
- Overall coherence and narrative flow
- Consistency of analogies throughout
- Readability and engagement
- No remaining gaps or confusions
- Ready for expert re-encoding?
**Prompt for thoughtbox:**
> "Final validation of [TOPIC] notebook. Will check overall coherence, consistency of analogies, readability, engagement, and confirm no gaps remain."
---
### Final Issues
*Last chance to catch problems:*
#### Coherence Issues
1.
2.
#### Consistency Problems
1.
2.
#### Readability Concerns
1.
2.
#### Remaining Gaps
1.
2.
---
### Action Items
- [ ] Polish issue 1
- [ ] Polish issue 2
- [ ] Final validation check
---
### Final Refinements
*Last touches:*
1. **Polished**: [Final change]
2. **Validated**: [Confirmation]
---
**β
Feynman Explanation is now complete and validated**
*Update the Feynman Explanation section above β for the last time*
---
# Phase 4: Expert Re-Encoding
*Now that understanding is validated, extract patterns for expert use.*
**Goal**: Create high-signal, low-friction reference material optimized for rapid application.
**Characteristics:**
- Pattern-dense with code examples
- Decision tables and lookup charts
- If-then logic, not narrative
- Anti-patterns explicitly marked (β)
- Checklists for validation
- ASCII diagrams for mental models
- Hierarchical for easy scanning
---
## Core Mental Model
*Extract the essence into ASCII diagrams:*
\`\`\`
[Create ASCII diagram showing system architecture or concept relationships]
Example:
βββββββββββ
β Client β
ββββββ¬βββββ
β
ββββββΌβββββ
β Server β
ββββββ¬βββββ
β
ββββββΌβββββ
βDatabase β
βββββββββββ
\`\`\`
**Key Principles:**
1.
2.
3.
---
## Pattern: [Pattern Name 1]
**When to Use:**
- Scenario 1
- Scenario 2
**Implementation:**
\`\`\`typescript
// β
CORRECT: Pattern shown
class Example {
// Show the right way
}
// β WRONG: Anti-pattern shown
class BadExample {
// Show what NOT to do
}
\`\`\`
**Trade-offs:**
| Aspect | Pro | Con |
|--------|-----|-----|
| | | |
---
## Pattern: [Pattern Name 2]
*Repeat structure for additional patterns*
**When to Use:**
- Scenario 1
- Scenario 2
**Implementation:**
\`\`\`typescript
// β
CORRECT: Pattern shown
// β WRONG: Anti-pattern shown
\`\`\`
**Trade-offs:**
| Aspect | Pro | Con |
|--------|-----|-----|
| | | |
---
## Decision Framework
*When should I use approach A vs B vs C?*
| Decision | Use When | Pros | Cons | Example |
|----------|----------|------|------|---------|
| Approach A | | | | |
| Approach B | | | | |
| Approach C | | | | |
---
## Common Scenarios
**Scenario 1: [Description]**
β Use: [Approach]
β Because: [Reasoning]
**Scenario 2: [Description]**
β Use: [Approach]
β Because: [Reasoning]
---
## β Anti-Patterns & Pitfalls
### Anti-Pattern 1: [Name]
**What it looks like:**
\`\`\`typescript
// Bad code example
\`\`\`
**Why it's wrong:**
- Reason 1
- Reason 2
**Do this instead:**
\`\`\`typescript
// Good code example
\`\`\`
---
### Anti-Pattern 2: [Name]
**What it looks like:**
\`\`\`typescript
// Bad code example
\`\`\`
**Why it's wrong:**
- Reason 1
- Reason 2
**Do this instead:**
\`\`\`typescript
// Good code example
\`\`\`
---
## Quick Reference Checklist
**Before using [TOPIC]:**
- [ ] Check 1
- [ ] Check 2
- [ ] Check 3
**Common gotchas:**
- β οΈ Watch out for X
- β οΈ Remember Y
- β οΈ Don't forget Z
**Resources:**
- [Link to docs]
- [Link to examples]
---
# Meta-Reflection
## What I Learned
**Core Understanding:**
[What is the main thing you now understand that you didn't before?]
**Surprises:**
[What was unexpected or counterintuitive?]
**Still Unclear:**
[What remains fuzzy? (Be honest)]
---
## Skill Export Checklist
- [ ] Create \`.claude/skills/[SKILL_NAME]/\` directory
- [ ] Write \`SKILL.md\` with expert re-encoding content
- [ ] Add code examples as separate files if needed
- [ ] Test skill by using it in a real scenario
- [ ] Archive this notebook as reference
---
## Success Validation
**Can you now:**
- [ ] Explain this topic clearly to a novice?
- [ ] Explain this topic precisely to an expert?
- [ ] Identify when to use vs not use this?
- [ ] Spot anti-patterns in real code?
- [ ] Implement this without constant reference to docs?
If all boxes are checked: **Understanding validated β
**`,
} as const;
export type TemplateName = keyof typeof TEMPLATES;
export const AVAILABLE_TEMPLATES: readonly TemplateName[] = [
'sequential-feynman',
] as const;
/**
* Get template content by name
* @throws Error if template not found
*/
export function getTemplate(name: string): string {
const template = TEMPLATES[name as TemplateName];
if (!template) {
throw new Error(`Template not found: ${name}. Available templates: ${AVAILABLE_TEMPLATES.join(', ')}`);
}
return template;
}