Practice by domain
The mock exam is for measuring where you stand. These pages are for the thing you do afterwards: drilling the one domain that came back weak, without the clock and without the other four getting in the way.
306 questions across the five domains, each with an explanation for every option rather than just the correct one — because on a judgement test, knowing why the other three fail is most of the skill.
- 27%Agentic Architecture & OrchestrationAgentic loops, coordinator-subagent orchestration, context passing, workflow enforcement, hooks, task decomposition, and session state.70 questions · 16 of the 60 on a real sitting come from this domain
- 18%Tool Design & MCP IntegrationTool interface design, structured error responses, tool distribution and tool_choice, MCP server integration, and built-in tool selection.50 questions · 11 of the 60 on a real sitting come from this domain
- 20%Claude Code Configuration & WorkflowsCLAUDE.md hierarchy, slash commands and skills, path-specific rules, plan mode, iterative refinement, and CI/CD integration.61 questions · 12 of the 60 on a real sitting come from this domain
- 20%Prompt Engineering & Structured OutputExplicit criteria, few-shot prompting, structured output via tool use, validation and retry loops, batch processing, and multi-pass review.65 questions · 12 of the 60 on a real sitting come from this domain
- 15%Context Management & ReliabilityContext window management, escalation and ambiguity, error propagation, codebase exploration, human review calibration, and provenance.60 questions · 9 of the 60 on a real sitting come from this domain