References: The Skills System

1. Modular programming - Wikipedia - Foundational coverage of decomposing software into independent modules; the conceptual ancestor of the Skills approach to bundling capability with lazy loading.

2. Lazy loading - Wikipedia - Detailed explanation of the lazy-loading pattern that makes Skills token-efficient; only the trigger description sits in context until the harness needs the body.

3. Plug-in (computing) - Wikipedia - Background on extensible software architectures where third-party capabilities load on demand; Skills are essentially a plug-in system for LLM agents.

4. AI Engineering - Chip Huyen - O'Reilly - Chapters on agent design and tool composition cover the broader patterns Skills implement, including task decomposition and capability binding.

5. Designing Machine Learning Systems - Chip Huyen - O'Reilly - The model-as-component framing in this book applies directly to Skills as composable units of model-augmented capability.

6. Anthropic Skills Documentation - Anthropic - Authoritative reference for the Skill format including frontmatter, body structure, bundled scripts, and the loading lifecycle covered in this chapter.

7. Anthropic Engineering: Building Skills for Claude - Anthropic - The announcement and design rationale for Skills; explains why the trigger-description-then-body architecture saves tokens at scale.

8. Claude Skills GitHub Examples - Anthropic GitHub - Official example Skills demonstrating the format conventions, script delegation patterns, and asset bundling used in this chapter's worked examples.

9. Model Context Protocol (MCP) - Anthropic - The protocol Skills use to expose capabilities to harnesses; relevant for engineers building custom Skills that interact with external systems.

10. Awesome Claude Skills - Dan McCreary GitHub - Curated collection of community-built Skills that demonstrate the trigger-description authoring patterns and script-delegation refactoring discussed in this chapter.