Study Vault - Pre-Note Generation Skill

Study Vault - Pre-Note Generation Skill

Argument Parsing

• No argument → start from source selection
• [book-name keyword] → start immediately with that book (PDF flow)
• [URL] → fetch and generate vault from the URL content (web/GitHub flow)
• [book-name] [chapter-range] → specific chapters only (e.g., kafka Ch3-Ch5)

Argument: $ARGUMENTS

---

Difference from /oh-my-study-with-me:study

	study	study-vault
Approach	Read and converse interactively to build understanding	Analyze the whole first, then generate notes
Speed	Slow (interactive)	Fast (auto-generated)
Depth	Deep (Why? conversation)	Broad (grasp overall structure)
Best for	Deep diving into key chapters	Bird's-eye view of a new book, structuring concept relationships, review notes

The two skills are complementary:

1. Use /oh-my-study-with-me:study-vault to get an overview of the book's entire structure first.
2. Use /oh-my-study-with-me:study to deep dive into key chapters.
3. Review using the vault's practice problems.

---

Design Principles

1. Source verification: Only write notes after actually reading and confirming the content, not just the filename.
2. Active recall: Write all answers in collapsed state (details/summary). Make the learner think before expanding.
3. Concept connections: Every note cross-references related notes.
4. Self-review: Run a quality checklist at the end.

---

Phase 1: Source Exploration and Structure Analysis

Detect the source type from the argument or user input:

Web URL Flow

1. Use WebFetch to retrieve the article/documentation content.
2. Extract the content structure (headings, sections, subsections).
3. Build a topic hierarchy from the extracted structure.
4. Continue to step 3 below.

GitHub Repo Flow

1. Use WebFetch to fetch the README.
2. Use gh CLI or WebFetch to get the repo structure (directory tree, key files).
3. Identify the main architectural components and documentation sections.
4. Build a topic hierarchy from the repo structure and README.
5. Continue to step 3 below.

Book/PDF Flow (Existing Behavior)

1. Read the PDF of the book the user selected.
   • If over 100MB, use pdftotext CLI (specify page range)
   • Extract the table of contents pages first (usually the first 5–15 pages).
2. Build a topic hierarchy from the table of contents:

   Part 1: Basics
   ├── Ch1: Introduction → [Topics: core concepts, architecture]
   ├── Ch2: Installation → [Topics: environment setup] (skip: replaced by lab)
   └── Ch3: Basic operations → [Topics: CRUD, internal structure]
   Part 2: Advanced
   └── ...
   

3. Show the topic hierarchy to the user and confirm the scope.
   • User can choose "generate all" vs "specific parts only"
   • Suggest skipping chapters like installation/environment setup, as they can be replaced with the /oh-my-study-with-me:lab skill

---

Phase 2: Content Analysis and Tag Design

1. Read each section from the source material:
   • PDF: Read each chapter in the selected scope, 20 pages at a time.
   • Web URL: Work with the fetched article sections.
   • GitHub repo: Analyze key source files, documentation, and architectural components.
2. Extract from each chapter:
   • Key concepts list (name + one-line definition)
   • Concept dependency relationships (A must be understood before B)
   • Real-world connections (where this concept is actually used)
3. Design tag standards:
   • Format: English kebab-case
   • Examples: #producer, #consumer-group, #replication, #partition
   • Show the tag list to the user and get approval.

---

Phase 3: Vault Creation

Directory Structure

study-vault/{category}/{book-name}/
├── 00-dashboard.md
├── 00-quick-reference.md
├── 00-concept-compare.md
├── 01-[topic-name]/
│   ├── concepts.md
│   └── practice.md
├── 02-[topic-name]/
│   ├── concepts.md
│   └── practice.md
└── ...

3-1. Dashboard (00-dashboard.md)

# [Book Name] Study Vault

## Learning Map

[Visualize topic dependency relationships in ASCII]

## Topic List

| # | Topic | # of Key Concepts | # of Practice Problems | Difficulty | Tags |
|---|-------|-------------------|------------------------|------------|------|
| 01 | [topic name] | N | N | ★☆☆ | #tag1, #tag2 |
| 02 | [topic name] | N | N | ★★☆ | #tag3, #tag4 |

## Weak Areas
_(Automatically reflected from the tracking/ metacognition file after studying)_

## References
- Source: [Book Name] ([Author], [Publisher])
- Created: YYYY-MM-DD
- Related skills: Deep dive with `/oh-my-study-with-me:study`, hands-on practice with `/oh-my-study-with-me:lab`

3-2. Quick Reference (00-quick-reference.md)

# Quick Reference

## Key Terms
| Term | Definition | Related Topic |
|------|------------|---------------|

## Key APIs / Commands / Configuration Values
| Item | Description | Default | Notes |
|------|-------------|---------|-------|

## Common Patterns
[Code snippets, configuration examples, etc.]

3-3. Concept Comparison (00-concept-compare.md)

# Comparison of Easily Confused Concepts

## [Concept A] vs [Concept B]

### Structural Differences
| Dimension | Concept A | Concept B |
|-----------|-----------|-----------|
| Problem it solves | ... | ... |
| How it works | ... | ... |
| When to use | ... | ... |
| Trade-offs | ... | ... |

### Why They Get Confused
[Structural explanation of why they appear similar on the surface]

### Key Distinguishing Criterion
[The essential difference that distinguishes them in one line]

### Related Concepts
- [Concept C](../03-topic/concepts.md#concept-c) is a higher-level abstraction of Concept A
- [Concept D](../04-topic/concepts.md#concept-d) is used together with Concept B

3-4. Concept Notes (concepts.md)

concepts.md in each topic folder:

# [Topic Name]

> Source: [Book Name] [Chapter Name] (p.XX-XX)
> Tags: #tag1 #tag2

## The Core Problem This Topic Solves
[One line describing the core problem this topic addresses]

## Overall Structure
[ASCII diagram of how the concepts in this topic fit together]

## Key Concepts

### [Concept 1]

**Definition**: ...

**Why it's needed**: [What problem arises without this concept. Explain its reason for existence structurally]

**How it works**:
[Include an ASCII diagram if helpful]

**Design trade-offs**: [What is gained and what is sacrificed by choosing this approach]

**Related concepts**: [Concept 2](../02-topic-name/concepts.md#concept-2), [Concept 3](#concept-3)

Rules:

• Each concept follows the order: Definition → Why it's needed → How it works → Design trade-offs → Related concepts
• "Why it's needed" is written as "what goes wrong without it." Explain its reason for existence, not a feature list
• "Design trade-offs" specifies the direction the concept chose and the cost that comes with it. Can be omitted if not applicable
• Visualize concept relationships with an "Overall Structure" diagram at the start of each topic
• Related concepts are cross-linked with in-file anchors or links to other topic files

3-5. Practice Problems (practice.md)

practice.md in each topic folder:

# [Topic Name] Practice Problems

> Difficulty: ★ Basic / ★★ Applied / ★★★ Advanced

## Problem 1 ★
[Question]

<details>
<summary>Show Answer</summary>

[Answer + explanation of why]

</details>

Rules:

• Minimum 8 problems per topic
• Difficulty distribution: Basic 40%, Applied 40%, Advanced 20%
• Type distribution: Recall 60%+, Application 20%+, Analysis 10%+
• Zero-hint principle: No answer hints must leak into the problem text
• All answers must be collapsed with <details><summary> tags to encourage active recall

---

Phase 4: Cross-Linking

1. Connect "Related concepts" in all concepts.md files bidirectionally.
2. Link from the topic list in the dashboard to each topic folder.
3. Link from concept-compare to related concepts.md files.
4. Link format: relative path markdown link [concept name](../01-topic/concepts.md#anchor)

---

Phase 5: Self-Review

After generation is complete, run the checklist below. If any item fails, fix it.

Coverage Review

• Are the key concepts of all selected chapters included in concepts.md?
• Does each topic have at least 8 practice problems?
• Are all major APIs/commands/configuration values included in quick-reference without omission?

Quality Review

• Does every concept have a "Why it's needed" section? (Fails if only a plain definition is present)
• Is the zero-hint principle upheld in all practice problems?
• Are all answers collapsed with <details> tags?

Link Review

• Do all cross-reference links point to actual files/anchors?
• Are all topics linked from the dashboard?

Source Review

• Is the source (book name, chapter, page) stated in all concepts.md files?
• Are only contents actually confirmed from the PDF written? (No guessing)

---

Integration with /oh-my-study-with-me:study

When deep studying with /oh-my-study-with-me:study after vault creation:

1. In Phase 1, reference the vault's concepts.md to supplement principle extraction.
2. Phase 3 verification results are recorded in the same tracking/ metacognition file.
3. Practice problems from the vault that were answered incorrectly can be automatically added to quiz_bank.json.

---

Notes

• Do not copy source content verbatim. Extract the essence and restructure from the learner's perspective.
• Read PDFs in 20-page chunks. Use pdftotext for PDFs over 100MB.
• For web sources, if the fetched content is insufficient, use WebSearch to supplement.
• For GitHub repos, focus on architecture, design decisions, and key implementations — do not dump entire codebases.
• Vault generation takes time. Report progress to the user step by step.
• If a vault already exists, confirm with the user whether to overwrite or append.