Skip to content

References: Using a Skill to Generate a Learning Graph

  1. Directed acyclic graph - Wikipedia - Covers topological ordering, cycle-free structure, and algorithms for detecting cycles, the exact mathematical property Automated Quality Validation checks in this chapter's dependency CSV before a graph is trusted.

  2. Human-in-the-loop - Wikipedia - Explains systems where a human reviews or corrects an automated process's output, the pattern this chapter formalizes as Human-in-the-Loop Review at checkpoints like Course Description Scoring and Concept List Review.

  3. Software agent - Wikipedia - Defines programs that act autonomously on a user's behalf with persistence and reactivity, the general category this chapter's Agent Skill and Claude Code sections narrow down to a filesystem-executing, skill-following coding agent.

  4. Building Intelligent Interactive Tutors - Beverly Park Woolf - Morgan Kaufmann - Surveys automated authoring pipelines and quality control for AI-generated educational content, paralleling this chapter's Concept List Generation, Taxonomy Generation Step, and Automated Quality Validation as a repeatable production pipeline.

  5. Introduction to Algorithms (4th Edition) - Thomas H. Cormen, Charles E. Leiserson, Ronald L. Rivest, Clifford Stein - MIT Press - The standard reference on graph algorithms, including cycle detection and topological sort, underlying the Automated Quality Validation step's checks for cycles, orphans, and disconnected components.

  6. Agent Skills - Anthropic - Official documentation defining a Skill as a reusable, filesystem-based instruction set an agent loads on demand, the precise architecture this chapter's Agent Skill section describes for the learning-graph-generator skill.

  7. Claude Code overview - Anthropic - Introduces Claude Code as an agentic tool that reads files, runs commands, and follows multi-step instructions inside a project, the host runtime this chapter names as required to execute any agent skill file.

  8. Extend Claude with skills - Anthropic - Explains how Claude Code discovers and triggers a SKILL.md file automatically or by direct invocation, illustrating this chapter's Skill Invocation section and its two triggering methods.

  9. Detect Cycle in a Directed Graph - GeeksforGeeks - Walks through DFS-based and topological-sort-based cycle detection with working code, the algorithmic technique underlying Automated Quality Validation's check that a dependency CSV forms a valid DAG.

  10. Best practices for Claude Code - Anthropic - Describes giving an agent a verifiable check, course-correcting early, and using a fresh subagent for adversarial review, patterns that mirror this chapter's Iterative Graph Regeneration and Human-in-the-Loop Review.