Practice
Learn at your own pace. Answer questions one at a time with instant feedback and explanations.
Start practiceClaude Certified Architect - Foundations
The Claude Certified Architect - Foundations (CCA-F) validates the ability to design and build production solutions with Claude across the Claude API, Claude Agent SDK, Claude Code, and the Model Context Protocol (MCP). It is scenario-based and covers agentic architecture and orchestration, tool design and MCP integration, Claude Code configuration and workflows, prompt engineering and structured output, and context management and reliability.
Mock exam
Simulate the real thing. Take a timed, full-length test and review your score and weak areas.
Sign up to startStudy your way: beyond Practice and Mock exam, choose adaptive, hard mode, ready review, objective coverage, or retry-your-misses — and set your own question count, timer, and pass mark.
About this exam
The Claude Certified Architect - Foundations exam assesses foundational, practical knowledge for building real-world solutions with Claude. It is scenario-based rather than trivia-focused and covers agentic systems, tool integration, Claude Code configuration, prompt engineering and structured output, and context management across the Claude API, Claude Agent SDK, Claude Code, and Model Context Protocol (MCP).
Who should take this exam
Developers, AI engineers, and solution architects who build agentic systems and applications on Claude. Anthropic recommends roughly six or more months of hands-on experience with the Claude API, Agent SDK, Claude Code, and MCP in production contexts.
Career benefits
Demonstrates the ability to architect reliable, production-grade Claude solutions — agents, tools, MCP integrations, and Claude Code workflows. Supports roles such as AI Engineer, Agent Developer, and Applied AI Architect.
How to prepare
Gain hands-on experience designing agentic loops and multi-agent systems, building tools with structured error handling, configuring MCP servers, configuring Claude Code (CLAUDE.md, rules, skills, slash commands, plan mode, CI/CD), prompt engineering and structured output with tool use and JSON schemas, the Message Batches API, and managing context and reliability.
Quick facts
Exam costVaries (see the Anthropic certification program)
Valid forPer the Anthropic certification program
Length90 minutes
Questions on exam60
Passing score720 / 1000
FormatScenario-based multiple choice (one correct answer with distractors).
Practice questions200
Objectives5
Official pageView
What's covered
1. Agentic Architecture & Orchestration
27%- 1a Design agentic loops and the agent execution model
- 1b Coordinate multi-agent systems and subagent spawning/delegation
- 1c Decompose tasks and apply orchestration patterns
- 1d Manage session and task state
2. Tool Design & MCP Integration
18%- 2a Design tools and structured error responses (errorCategory, isRetryable)
- 2b Configure and integrate MCP servers
- 2c Distribute tools across agents and manage tool access
3. Claude Code Configuration & Workflows
20%- 3a Configure Claude Code (CLAUDE.md hierarchy, .claude/ rules, settings)
- 3b Build custom skills, slash commands, and subagents
- 3c Use plan mode vs direct execution and manage permissions
- 3d Run Claude Code in CI/CD (headless mode, JSON output)
4. Prompt Engineering & Structured Output
20%- 4a Write effective prompts (explicit criteria, few-shot, system prompts)
- 4b Produce structured output with tool use and JSON schemas, with validation-retry
- 4c Process work at scale with the Message Batches API
- 4d Apply multi-pass review and evaluation
5. Context Management & Reliability
15%- 5a Preserve critical context across long conversations
- 5b Design escalation patterns and graceful failure handling
- 5c Manage error propagation and reliability in multi-agent setups
Frequently asked questions
Are these real exam questions?
No. These are original practice questions written to match the exam objectives, each with an explanation so you actually learn the material — not exam dumps.
How does practice mode work?
You answer questions one at a time with instant feedback and explanations. Over time the app adapts, prioritizing the objectives and questions you struggle with most.
What is a mock exam?
A timed, full-length simulation that holds feedback until the end, then shows your score, pass/fail result, and a breakdown by objective.
Can I customize how I study?
Yes. Pick the study mode that fits — adaptive practice, hard mode, ready-for-review, objective coverage, or retrying questions you've missed — and set your own question count, timer, and passing score for each session.
Do I need an account?
You can try free questions for this exam without signing in. Create a free account to save your progress, track weak objectives, and unlock the full question bank.