Curriculum
One study-notes page per exam domain, in exam weighting order. Read them top to bottom on a first pass, then come back to the heaviest domains. Each page ends with a pointer to the matching drill in practice mode.
- 27% Agentic Architecture & Orchestration 54 drill questions
Single calls vs workflows vs agents, the agentic loop, subagents and orchestration, cost/latency trade-offs, evaluation and recovery.
- 18% Tool Design & MCP Integration 36 drill questions
Tool schemas and descriptions, tool_choice, handling results and errors, Model Context Protocol primitives and transports.
- 20% Claude Code Configuration & Workflows 40 drill questions
CLAUDE.md memory, settings hierarchy, permissions, hooks, slash commands, subagents, MCP servers, headless usage.
- 20% Prompt Engineering & Structured Output 40 drill questions
System prompts, XML structure, few-shot examples, long-context placement, JSON schema outputs, strict tool use.
- 15% Context Management & Reliability 30 drill questions
Prompt caching mechanics, token counting, batching, rate limits and error handling, compaction, output limits.
Independent community study resource — not affiliated with or endorsed by Anthropic. Claude is a trademark of Anthropic, PBC. All study material and practice questions here are original, written from Anthropic's public documentation. Everything runs in your browser; nothing you answer is stored or transmitted.