Skip to content

Learning Graph: Concept Map Overview

This page documents the complete concept dependency graph for the Custom Skill Developer knowledge base. The graph contains 200 concepts spanning all 17 chapters, organized by category, difficulty, and Bloom's Taxonomy level. Each concept is specific, testable, and linked to its prerequisite concepts via the dependencies column in concepts.csv.


Concept Inventory Summary

Metric Count
Total concepts 200
Categories 12
Difficulty levels 3
Bloom's levels 5
Chapters covered 16 of 17

Breakdown by Category

Category Concept Count Description
Pipeline Orchestration 32 DAG design, checkpoints, file contracts, state management, orchestrator patterns
Meta-Skills 25 Routing tables, decision trees, references directory, disambiguation protocols
Specialized Skills 23 Data format, code generation, analysis, and validation skill types
Deployment 22 Installation paths, symlink patterns, registry, distribution, versioning
Skill Anatomy 20 SKILL.md sections, workflow steps, development process, iteration loop
Foundations 17 Core definitions, execution model, skill categories, background concepts
Token Efficiency 12 Lazy loading, progressive disclosure, skip-if-complete, tiered retrieval
Workflow Design 11 Step patterns, branching, user dialog triggers, error handling
Quality Systems 10 Rubric design, scoring, thresholds, gap reports, pipeline integration
Testing 10 Manual testing, debugging, branch testing, edge cases
Ecosystem 9 Skill registry, discovery, 30-skill limit, taxonomy, distribution
YAML Specification 9 Required and optional frontmatter fields and their constraints

Breakdown by Difficulty Level

Difficulty Concept Count Coverage
Beginner 46 Foundational vocabulary, core definitions, first principles
Intermediate 144 Applied patterns, design decisions, multi-concept integration
Advanced 10 Pipeline coordination, parallel execution, orchestrator design

The distribution reflects the knowledge base's audience: intermediate-to-advanced Claude Code users who need precise, actionable knowledge at the applied level.


Breakdown by Bloom's Taxonomy Level

Bloom's Level Concept Count Example Concepts
Remember 33 SKILL.md file format, name field requirement, 30-skill limit
Understand 73 Skill execution model, meta-skill routing, lazy loading definition
Apply 65 Routing table construction, quality rubric design, symlink installation
Analyze 15 Routing failure modes, common skill failure modes, critical path identification
Evaluate 14 Production-readiness checklist, meta-skill anti-patterns, escape hatch rate
Create 0 Covered implicitly through applied project work across all chapters

The heavy concentration at the understand and apply levels reflects the knowledge base's emphasis on building functional skills over memorizing theory.


Concept Count by Chapter

Chapter Title Concepts
1 What Are Claude Code Skills? 27
2 The Skill Ecosystem 8
3 Your First Skill 12
4 SKILL.md Structure 12
5 YAML Frontmatter
6 Workflow Design 4
7 Quality Scoring Systems 9
8 Meta-Skill Routers 25
9 Token Efficiency Design 11
10 Session Logging 7
11 Pipeline Orchestration 29
12 Data Format Skills 6
13 Code Generation Skills 9
14 Analysis & Validation Skills 8
15 Installation & Registry 18
16 Testing & Debugging 9
17 Publishing & Distribution 6

Note: YAML Specification concepts (Chapter 5) are attributed to Chapter 1 where they first appear structurally, since frontmatter is introduced with the first SKILL.md definition.


Dependency Graph Structure

Concepts build on each other in a directed acyclic graph (DAG). The root node is C001: Claude Code skill definition — all other concepts ultimately depend on it. Key dependency chains:

Foundations chain: C001 → C002 → C027 → C028/C029/C031 (skill → file → frontmatter → fields)

Workflow chain: C001 → C007 → C047 → C048/C053 → C055/C056 (skill → execution → steps → patterns → error handling)

Quality chain: C002 → C043 → C060 → C062/C063 → C065 → C069 (skill → section → rubric → scoring → gate → pipeline gate)

Meta-skill chain: C001 → C076 → C082 → C083/C087 → C091/C092 → C093 (skill → meta-skill → routing → table/tree → references → lazy loading)

Pipeline chain: C001 → C121 → C122 → C127 → C128/C133 → C137/C138 (skill → pipeline → stage → DAG → checkpoints → state/contracts)


How to Use This Graph

The CSV file concepts.csv contains the following columns:

Column Description
id Unique concept identifier (C001–C200)
concept_name Human-readable concept label
category Top-level category (12 total)
subcategory Finer grouping within the category
difficulty_level beginner / intermediate / advanced
bloom_level remember / understand / apply / analyze / evaluate
dependencies Comma-separated list of prerequisite concept IDs
chapter Primary chapter where the concept is introduced

Use this data to:

  • Sequence learning — follow dependency chains from C001 outward
  • Assess readiness — verify prerequisite concepts before introducing advanced ones
  • Generate quizzes — filter by Bloom's level to create targeted assessments
  • Visualize curriculum — import into graph visualization tools (vis-network, D3.js) using the id and dependencies columns as edges
  • Identify coverage gaps — compare chapter concept counts to ensure balanced coverage

Data File

The raw concept data is stored at:

docs/learning-graph/concepts.csv

The file is UTF-8 encoded, comma-delimited, with a header row. Multi-valued dependency fields use comma-separated values enclosed in double quotes per standard CSV conventions.