Anthropic Claude Certified Architect - Foundations Exam Prep

Covers Claude Code setup, configuration hierarchy, command usage, automation hooks, subagents, and team workflow integration. This domain emphasizes how to configure Claude Code safely and effectively across user, project, and local scopes, and how to orchestrate work with commands, hooks, agents, and IDE/MCP integrations. Candidates should be ready to connect these concepts to realistic architecture and workflow decisions, especially where configuration scope, tool boundaries, context management, and operational reliability affect production use.

• Claude Code installation, configuration, and project initialization
• Slash commands, hooks, and automation
• Subagents and Task tool orchestration
• IDE integration, MCP servers, and team workflows
• Settings hierarchy (user/project/local), permission modes, plan mode
• /init, /clear, /resume, /compact commands and when to use each
• Project CLAUDE.md vs user CLAUDE.md inheritance and override semantics

Key references: CCA-F outline Domain 1 · ExamPal shared topic tree

Covers agent design patterns, orchestration strategies, asynchronous execution, handoff workflows, and core Agent SDK concepts. This domain focuses on choosing the right agent structure, coordinating work across agents, and managing lifecycle and context flow in agentic systems. Candidates should be ready to connect these concepts to realistic architecture and workflow decisions, especially where configuration scope, tool boundaries, context management, and operational reliability affect production use.

• Single-agent vs multi-agent design patterns
• Background work, async patterns, and notifications
• Multi-step workflows, enforcement, and handoff
• Agent SDK fundamentals
• Coordinator–worker, supervisor–subagent, peer-to-peer patterns
• When to spawn subagents (parallel research, blast-radius isolation, context protection)
• Task decomposition strategies

Key references: CCA-F outline Domain 2 · ExamPal shared topic tree

Covers prompting strategies, structured output techniques, long-context optimization, and reliability methods for Claude outputs. This domain emphasizes how to shape model behavior, produce valid machine-readable output, and verify results through sampling, critique, and grounding. Candidates should be ready to connect these concepts to realistic architecture and workflow decisions, especially where configuration scope, tool boundaries, context management, and operational reliability affect production use.

• Prompting techniques for Claude
• Structured output and JSON mode
• Long-context, prompt caching, and cost optimization
• Output reliability and verification
• Few-shot placement (top of prompt vs interleaved)
• Zero-shot, chain-of-thought, ReAct-style prompting
• System vs user message roles, prefilling assistant turns

Key references: CCA-F outline Domain 3 · ExamPal shared topic tree

Covers tool interface design, MCP server integration, structured error handling, and tool selection behavior. This domain focuses on building reliable tools and integrating them safely through MCP with appropriate trust, authentication, and error semantics. Candidates should be ready to connect these concepts to realistic architecture and workflow decisions, especially where configuration scope, tool boundaries, context management, and operational reliability affect production use.

• Tool interface design
• MCP server design and integration
• Structured error responses
• Tool selection and tool-use loops
• Tool descriptions, input boundaries, parameter clarity
• Naming conventions, idempotency declarations
• Tool documentation as instruction-following surface area

Key references: CCA-F outline Domain 4 · ExamPal shared topic tree

Covers context window management, memory architectures, production reliability patterns, and observability/debugging practices. This domain emphasizes preserving useful context, managing memory across sessions, and operating agents reliably with strong monitoring and testing. Candidates should be ready to connect these concepts to realistic architecture and workflow decisions, especially where configuration scope, tool boundaries, context management, and operational reliability affect production use.

• Context window economics
• Memory architectures
• Reliability patterns for production agents
• Observability and debugging
• Token budgeting, sliding windows, summarization
• Compaction triggers, lossy vs lossless reduction
• File-system memory for persistence beyond a session

Key references: CCA-F outline Domain 5 · ExamPal shared topic tree