Take 60 original questions I wrote while studying for Claude Certified Architect Foundations. Choose study mode or a timed 120-minute attempt. Your results show which domains need more work, and this browser saves your progress so you can return later. The questions are in English. The test is free and needs no account.
Updated
This mock uses single-answer questions across all six scenarios. The official CCAR-F also has multiple-response items and selects four of six scenarios. The mock uses its own scoring formula, with no validated correspondence to Anthropic’s exam score.
Full Practice Exam · 60 Questions
Retake guarantee. Score 720+ here in exam mode and send the result from the email address you pay with: fail the real Architect exam as a member and your next 2 months are free.
This mock uses its own formula: 100 + 900 × (correct ÷ total). Its 720 threshold is a practice reference, not Anthropic’s scoring method or a validated prediction of your official result.
Unofficial practice material. This test is not authorized or endorsed by Anthropic. The original questions are based on the publicly available Claude Certified Architect Foundations exam guide.
The questions in this section come from the same bank as the test above. The domain weights are from Anthropic’s Claude Certified Architect Foundations exam guide (v1.0, July 2026).
| Domain | Weight in the exam guide | Questions in this mock |
|---|---|---|
| D1 Agentic Architecture & Orchestration | 27% | 16 |
| D2 Tool Design & MCP Integration | 18% | 11 |
| D3 Claude Code Configuration & Workflows | 20% | 12 |
| D4 Prompt Engineering & Structured Output | 20% | 12 |
| D5 Context Management & Reliability | 15% | 9 |
Pick an answer before you open the explanation. Every question in the test above works the same way.
The synthesis subagent's reports read as generic overviews: in 30% of runs they omit the specific figures and named sources the web-search subagent surfaced. Tracing shows the Task prompt sent to the synthesis subagent contains only the report topic and formatting instructions. What is the best fix?
Correct answer: B. Subagents do not inherit the coordinator's conversation context; in a hub-and-spoke topology the coordinator must pass prior findings explicitly in the Task prompt, so embedding the web-search and document-analysis results verbatim is the fix. A larger context window (A) does not cross the process boundary into a freshly spawned subagent. Instructing it to 'recall' earlier research (C) fails because the subagent starts with no prior turns to recall. Giving it its own search tool (D) is wrong scope and over-engineering — it duplicates work and diverges from the evidence already gathered.
The document-analysis subagent must call classify_document before running any field extraction, because the extraction schema depends on the detected type. Relying on a prompt instruction, it skips classification in 9% of runs and extracts against the wrong schema. The coordinator invokes it once per document, and the required first step is always the same. What is the best fix?
Correct answer: B. When a specific tool must run first deterministically, forcing tool_choice to that named tool on the opening turn is the direct API mechanism and guarantees classify_document executes before any extraction. Setting tool_choice "any" (A) forces a tool but not which one, so the model could still call an extraction tool. A stronger prompt (C) is exactly the non-deterministic approach that already failed 9% of the time. An interception hook (D) would work but is heavier cross-tool policy machinery for a single fixed first-step ordering that the forced tool_choice already guarantees at the API level.
Your team standardized naming and error-handling conventions by adding them to ~/.claude/CLAUDE.md on the tech lead's machine. Everything works for the lead, but three newly onboarded engineers who cloned the repo report that Claude Code ignores the conventions entirely, even though `git status` shows a clean checkout. What is the most effective fix?
Correct answer: C. The CLAUDE.md hierarchy loads user-scoped instructions from ~/.claude/CLAUDE.md, which is personal to a machine and is NOT shared through version control — the classic 'new teammate didn't get the instructions' symptom. Committing the conventions to the project-scoped .claude/CLAUDE.md (or repo root) makes them load for every clone, so C is correct. An @import to ~/.claude/CLAUDE.md (A) fails because that file doesn't exist on the teammates' machines. Running /memory (B) only inspects already-loaded memory files; it cannot pull in instructions that were never in a shared scope. A skill (D) is on-demand rather than always-loaded, so it wouldn't provide the persistent conventions the team wants.
No. This is an independent mock exam with 60 original questions written from Anthropic’s public Claude Certified Architect Foundations exam guide. It is not affiliated with or endorsed by Anthropic, and no question is taken from the exam itself.
Anthropic’s exam guide (v1.0, July 2026) lists 60 multiple-choice and multiple-response questions, a 120-minute time limit and a pass at a scaled score of 720 on a 100 to 1,000 scale.
The exam guide weights the 5 domains as follows: Agentic Architecture & Orchestration 27%; Tool Design & MCP Integration 18%; Claude Code Configuration & Workflows 20%; Prompt Engineering & Structured Output 20%; Context Management & Reliability 15%. The questions in this mock belong to the same domains, so your results show where to study.
The score is 100 + 900 × the share of correct answers, and 720 is the threshold. The formula is mine, not Anthropic’s, so the result is a practice reference and does not predict your exam score.
No. The test is free and runs in your browser, which saves your progress so you can come back later. Your score and every explanation appear without an email address.
Add szymonpaluch.com as a connector in Claude to work through the same 60 questions in chat. You get one question at a time. Pick your answer, then see the correct choice and its explanation. No signup or API key is required.
https://szymonpaluch.com/mcp claude mcp add --transport http szymonpaluch https://szymonpaluch.com/mcp Questions are in English. The same server can also search and read this blog, list my services, prices and talks, and give you the booking link. It works with any MCP client that supports remote (Streamable HTTP) servers.
Want to train the team you write code with? The scope and prerequisites are below. Send this section’s link to the person who approves the budget.
Describe what you want to build with your team on the workshop page. I will reply to your enquiry.