What the Claude Certified Architect – Foundations exam actually asks.

• 60 questions in 120 minutes — exactly 2 minutes each on an even split. The prompts are scenario-based, so the pace comes from reading, not arithmetic.
• It sits at the Foundations level, the entry point of the Claude certification path, and assumes you build with Claude rather than train models.
• All 60 are scored. Unlike exams that seed unscored trial items, every question here counts.
• 720 out of 1000 to pass, on a scaled score. Because it is scaled, the raw number of correct answers you need is not fixed or published.
• Question formats: multiple choice and multiple response. A multiple-response item tells you how many to select and expects exactly that many.
• Answer everything. A blank scores nothing, so flag the long ones, bank the quick ones, and return before time runs out.
• Available to the Claude Partner Network. The exam is offered to member organisations rather than to anyone; eligibility sits with your employer's partner status, not with you individually. Confirm your organisation's access before you plan a date.

Two hours is enough for sixty questions only if you keep moving. Decide early how long a scenario may take before you flag it and move on.

• Domain 1 — Agentic Architecture & Orchestration (27%). The largest domain. Designing agentic loops, deciding when the model or the program ends a turn, and coordinating multiple agents where one drives and the others return results.
• Domain 2 — Tool Design & MCP Integration (18%). Deciding what becomes a tool, writing the description the model selects on, and building and connecting MCP servers for tools, resources and prompts.
• Domain 3 — Claude Code Configuration & Workflows (20%). Project and user configuration, instruction files, hooks for deterministic behaviour, subagents, and running Claude Code in CI and headless workflows.
• Domain 4 — Prompt Engineering & Structured Output (20%). Explicit criteria, few-shot and worked examples, system prompts, and producing output a program can consume reliably through tool schemas and structured formats.
• Domain 5 — Context Management & Reliability (15%). What survives a long conversation, when to stop and ask, how failure travels between agents, and keeping a person in the loop with confidence and provenance.

Domains 1 and 3 together make up 47% of the exam. The smallest domain is 15%, which is still nine questions, so it is not safe to skip.

Expect a short scenario — an agent, a workflow, a configuration — followed by a question about the best design or the next change. More than one option may work in isolation; the exam asks for the one that satisfies every stated constraint, and the guide names the production situations the items are drawn from: agentic customer support, multi-agent research, Claude Code in CI, developer productivity tooling, and structured extraction. Anthropic publishes the exact task statements in the exam guide, so confirm them there before you sit the paper.

Read the last sentence first, then the scenario, then the options. Qualifiers to look for:

• "Deterministic" → Prefer a hook or a program check over asking the model to remember.
• "Reliably" or "a program must consume it" → Prefer a schema the model is constrained by over parsing prose.
• "Across many turns" → Ask what must survive summarisation and where it should be written down.

The questions test judgement about tradeoffs, so reading about features is not enough. The study that pays is applying that knowledge to scenarios and being able to say why one option fails a stated constraint.

Three habits that work:

• Build a small agent end to end. A tool-using loop with a hook, a clear tool description, and structured output touches most of the domains and makes the questions concrete.
• Read the Anthropic documentation alongside the guide. The Claude API, the Agent SDK, Claude Code, and MCP each have their own reference; the guide names the topics but the documentation settles the detail.
• Review every option, not just your answer. On this exam several options are plausible; learning why one breaks a specific constraint is the skill being tested.

MockRoom's CCAR-F study room has five distinct 60-question mock exams — 300 original scenario questions in total, written against the published exam guide and weighted across its five domains, each with an explanation for every option. It also has 95 active recall flashcards for spaced review. Nothing here is taken from an exam dump or from the guide's own sample questions, and none of it is real examination content.

Open the CCAR-F study room