Claude Certified Architect Practice Test | CCA Prep
Independent exam prep · Not affiliated with, authorized by, or endorsed by Anthropic.✕
×<br>Log In
Email Address
Password
Confirm Password
Forgot password?<br>Log In<br>Already have an account? Log In
Don't have an account? Sign Up Free
Please wait...
🎉
Welcome to Claude Certified Architects!
Your free account is ready. Here's where most people start:
See pricing & enroll — $49<br>Take the free diagnostic first ↗
Independent CCA Foundations practice tests
Claude Certified Architect Practice Test
Prepare for the CCA Foundations exam with realistic, scenario-based questions designed to match the exam format. Study all 5 exam domains and pass with confidence.
Start the Free Diagnostic →<br>View Lesson Plan
400<br>Practice questions
Exam domains
Test modes
Lifetime<br>Free updates
Already sure you want full access?<br>Enroll Now — $49
Built to Get You Exam-Ready
A complete, independent prep program covering all five CCA Foundations domains — every question written to match the official scenario-based format.
400
Scenario-based questions
100%
Of answers with written explanations
5/5
Exam domains covered, at official weightings
120 min
Full-length timed mock exam
Built from the official CCA Foundations exam guide · Updated weekly · 10-day money-back guarantee
for full reference).<br>Remove this comment and uncomment the wrapper below when ready.<br>████████████████████████████████████████████████████████████████████ -->
-->
See the question quality before you buy
Real scenario-based questions, every answer fully explained — just like the 400 in the full bank.
Agentic Architecture & Orchestration (27%)<br>A coding agent calls tools inside a loop. A developer needs to decide when to terminate the loop. Which is the correct, robust termination signal?
AStop when the assistant’s text no longer contains a completion word like “done.”
BStop after a fixed maximum of 10 iterations, regardless of state.
CKeep executing tools and looping while the response’s stop_reason is tool_use; stop when stop_reason is end_turn.
DStop as soon as any tool returns an error.
Show answer
✓ Correct: C
The API reports why generation stopped via stop_reason. tool_use means run the tools, append results, and loop back; end_turn means Claude is finished, so the loop ends. Parsing text for completion words is a brittle anti-pattern. A fixed iteration cap is a safety backstop, not the main control. Stopping on the first tool error prevents the agent from recovering.
Claude Code Configuration & Workflows (20%)<br>A team adds Claude Code to a CI pipeline to auto-review pull requests. The job hangs indefinitely and times out instead of producing output. Which fix resolves it?
ASet CLAUDE_HEADLESS=true.
BRun claude with the -p (print) flag and pass the prompt as an argument.
CAdd the --batch flag.
DRedirect stdin from /dev/null.
Show answer
✓ Correct: B
Without -p, claude starts an interactive session and waits on stdin for input that never comes, so it hangs. The -p / --print flag runs headless: take the prompt, print the result to stdout, exit. CLAUDE_HEADLESS and --batch are not real flags. Redirecting stdin from /dev/null is a hack that doesn’t give the print-and-exit behaviour you actually want.
Tool Design & MCP Integration (18%)<br>You’re designing a tool that queries an external inventory API for an agent. The API sometimes returns “item not found” or rate-limit errors. How should the tool handle these to maximise the agent’s ability to recover?
ARaise an exception that terminates the agent loop so a human can intervene.
BReturn a structured result describing the error (clear message and error type) as the tool result, so Claude can read it and decide how to proceed.
CReturn an empty string so Claude assumes there’s no data.
DRetry silently inside the tool until the API eventually succeeds.
Show answer
✓ Correct: B
Tool results are fed back to Claude, so a clear structured error lets it retry, try an alternative, or report back. Terminating on every error removes self-correction. An empty string hides the failure and invites hallucination. Silent infinite retries can hang the agent and hammer a rate-limited API.
COMPLETE COURSE + PRACTICE TESTS
$49.00
One-time payment · Lifetime access
Currency:
🇺🇸 USD — US Dollar<br>🇪🇺 EUR — Euro<br>🇬🇧 GBP — British Pound<br>🇨🇦 CAD — Canadian Dollar<br>🇦🇺 AUD — Australian Dollar<br>🇯🇵 JPY — Japanese Yen<br>🇨🇳 CNY — Chinese Yuan<br>🇮🇳 INR — Indian Rupee<br>🇧🇷 BRL — Brazilian Real<br>🇰🇷 KRW — Korean Won<br>🇲🇽 MXN — Mexican Peso<br>🇨🇭 CHF — Swiss Franc<br>🇸🇪 SEK — Swedish Krona<br>🇸🇬 SGD — Singapore Dollar<br>🇳🇿 NZD — New Zealand Dollar<br>🇿🇦 ZAR — South African Rand<br>🇳🇬 NGN — Nigerian Naira<br>🇦🇪 AED — UAE Dirham<br>🇸🇦 SAR — Saudi Riyal<br>🇵🇱 PLN — Polish Zloty<br>🇹🇭 THB — Thai Baht<br>🇮🇩 IDR — Indonesian Rupiah<br>🇵🇭 PHP — Philippine Peso<br>🇹🇼 TWD — Taiwan Dollar<br>🇭🇰 HKD — Hong Kong Dollar<br>🇮🇱 ILS — Israeli Shekel<br>🇨🇴 COP — Colombian...