Learn

Domain 1 established the lesson template the rest reuse, and was built and reviewed the slow way — one domain, studied first. Domains 2–5' lessons and practice questions are Claude-drafted at the owner's direction, pending a review pass; see the project's roadmap for exactly what that means per domain.

This is a sequence, not a gate: every domain is open, whether or not its content has been reviewed yet.

1. 1

   Agentic Architecture & Orchestration

   27%

   Agent loops, agent vs. deterministic workflow, single- vs. multi-agent design, delegation and subagents, parallelization, human-in-the-loop, stopping conditions, and the reliability trade-offs each choice carries.

   Open the lesson →
2. 2

   Tool Design & MCP Integration

   18%

   MCP clients and servers, tools vs. resources vs. prompts, tool schema design, permissions and boundaries, authentication, error handling, and designing safe MCP integrations.

   Open the lesson →
3. 3

   Claude Code Configuration & Workflows

   20%

   CLAUDE.md and project instructions, permissions, hooks, commands and skills, MCP inside Claude Code, context management, and development workflows.

   Open the lesson →
4. 4

   Prompt Engineering & Structured Output

   20%

   System and user prompts, few-shot examples, structured (XML/JSON/schema) prompting, tool-use prompting, validation and retries, and separating deterministic application logic from model reasoning.

   Open the lesson →
5. 5

   Context Management & Reliability

   15%

   Context windows, context engineering, summarization and compaction, retrieval, caching, error handling and retries, observability, evaluations, and production reliability.

   Open the lesson →