Beyond the Fundamentals

Claude Code Agents: Fundamentals covered the mental model behind a single agent — what it is, how it's defined, how to brief it well — and applied that model to seven distinct agent types, one at a time. This course assumes all of that is comfortable working knowledge, and goes further into what happens once agents stop being used one at a time: building them programmatically, chaining them together, and running them at real production scale.

A Quick Recap of What Fundamentals Covered

• The core mental model — subagents vs. the default assistant, tool-permission isolation, agent-definition anatomy (Chapters 1–2)
• Communicating with an agent — writing a self-contained brief, foreground vs. background invocation, reading a final report critically (Chapter 3)
• Seven individual agent types — teaching, coding, code-review, testing, refactoring, documentation, and research (Chapters 4–10)
• A capstone applying all seven to one real project, one at a time, by hand (Chapter 11)

If any of that feels shaky rather than second nature, it's worth revisiting those chapters directly — this course builds on each of them without re-teaching them.

What This Course Adds

• Building agents programmatically with the Claude Agent SDK — moving beyond hand-written agent definitions (Chapter 2)
• Parallel vs. sequential orchestration — running several agents at once versus one after another, and when each is right (Chapter 3)
• Background & autonomous agents — long-running and self-scheduling work (Chapter 4)
• Specialized agent types — security/compliance review agents, and planning/architecture agents (Chapters 5–6)
• Agent-to-agent handoffs & shared state — passing results and context directly between agents (Chapter 7)
• Debugging and evaluating agent behavior — what to do when an agent gets something wrong (Chapter 8)
• Cost, context & efficiency tradeoffs — when orchestration is actually worth its own overhead (Chapter 9)
• Capstone — a production-style multi-agent pipeline built for the `website-content` project itself, formalizing Chapter 11's own capstone into a real, chained pipeline (Chapter 10)

Who This Course Is For

Anyone comfortable with Claude Code Agents: Fundamentals' own eleven chapters — the mental model, the seven agent types, and briefing an agent well. No new software is required beyond what Fundamentals already assumed; this course goes deeper into coordination and scale, not into new prerequisite tools.

Hands-On Exercises