Study Domains
Five exam domains covering Claude architecture, tools, configuration, prompting, and reliability.
Design and implement agentic systems using the Claude Agent SDK, including multi-agent coordination, lifecycle management, and programmatic enforcement patterns.
Design effective tools, implement structured error handling, configure MCP servers, and optimize tool distribution across agent architectures.
Configure Claude Code for team workflows using CLAUDE.md hierarchy, custom rules, slash commands, skills, and CI/CD integration patterns.
Apply advanced prompt engineering techniques including structured output with JSON schemas, validation-retry loops, batch processing, and multi-pass review architectures.
Manage context windows effectively, handle error propagation across multi-agent systems, implement escalation patterns, and design human review workflows.