The Claude Certified Architect – Foundations exam (CCAR-F) does not test whether you can call an API. It tests whether you can design a system around Claude that still behaves when a tool fails, a subagent gets the wrong context, a prompt is ignored or a long session runs out of room. This course walks through every one of the 30 task statements in the official exam guide, in the guide’s own order and weighted by the exam blueprint, so you know what each one asks and why the right answer is right. Each lesson is a short narrated video, explained the way a tutor would rather than read from slides, with animated diagrams synchronised to the voice and English subtitles. Identifiers such as stop_reason, allowedTools, PreToolUse or tool_choice appear on screen exactly as you will meet them in the docs and in code.
What is covered:
- Domain 1 · Agentic architecture and orchestration (27%): agentic loops, coordinator and subagent patterns, context passing and spawning, enforcement and handoff, Agent SDK hooks, task decomposition, sessions, resumption and forking.
- Domain 2 · Tool design and MCP integration (18%): tool descriptions, structured error responses, tool distribution and tool_choice, MCP server configuration, built-in tools.
- Domain 3 · Claude Code configuration and workflows (20%): the project memory file hierarchy, custom commands and skills, path-specific rules, plan mode versus direct execution, iterative refinement, Claude Code in CI pipelines.
- Domain 4 · Prompt engineering and structured output (20%): explicit criteria, few-shot prompting, structured output through tool use and JSON schemas, validation and retry loops, batch processing, multi-instance and multi-pass review.
- Domain 5 · Context management and reliability (15%): preserving critical context, escalation and ambiguity, error propagation across agents, large-codebase exploration, human review and confidence calibration, provenance and uncertainty.
