Skip to content

Agentic Architecture & Orchestration Questions

Practice questions for Agentic Architecture & Orchestration topic in Claude Certified Architect – Foundations. 52 questions covering this domain.

52 questions10 easy28 medium14 hard
Q1
hard

A hub-and-spoke system's "Policy Lookup" subagent consistently returns irrelevant policy sections, even though the correct policy area was identified ...

Q2
medium

In a hub-and-spoke multi-agent architecture, what is the primary function of the coordinator agent?

Q3
medium

An engineer builds an agentic loop that exits when Claude's last message contains the phrase "Task complete." Why is this approach considered an anti-...

Q4
medium

A developer spawns a Claude subagent to process a customer refund request. The subagent fails to apply the correct refund policy even though that poli...

Q5
hard

A tool in an agentic loop throws a transient network error. What is the correct pattern for handling this error within the loop?

Q6
medium

A development team sets an agentic loop's iteration cap at 10 and uses it as the primary termination mechanism. The loop exits after 10 iterations eve...

Q7
easy

In the Claude API, which `stop_reason` value signals that the model has completed its response and the agentic loop should exit?

Q8
medium

Which `stop_reason` value indicates that the Claude API response was truncated because it reached the maximum token limit?

Q9
hard

A solution architect is designing a system that extracts key clauses from contracts, checks them against compliance rules, and generates a summary rep...

Q10
medium

A production system needs to retrieve data from three independent external APIs simultaneously to build a consolidated report. Which pattern best achi...

Q11
easy

What field in the Claude API response is the correct programmatic signal for whether an agentic loop should continue or terminate?

Q12
hard

A multi-agent Claude system runs indefinitely on certain customer queries, never terminating. The developer confirms that `stop_reason` is being check...

Q13
medium

An agentic session was interrupted when the server crashed mid-task. The session state was persisted before the crash. Which Claude Code session manag...

Q14
hard

A hub-and-spoke system coordinates a legal document analysis workflow. A subagent tasked with checking compliance regulations consistently returns irr...

Q15
medium

An AI application needs to answer a single-turn question by looking up a customer's account balance. Which architecture is most appropriate?

Q16
medium

When invoking a subagent in a hub-and-spoke system, what information must the coordinator explicitly provide?

Q17
hard

An agentic Claude Code session is partway through a complex database migration task when the developer must stop for the day. The developer wants to c...

Q18
hard

An agentic loop checks whether to terminate by evaluating if Claude's most recent message contains the text "I have completed all tasks." During testi...

Q19
medium

In a hub-and-spoke multi-agent system, a research coordinator delegates a "financial analysis" task to a subagent. What role does the coordinator play...

Q20
hard

A production agentic system processes 100 customer service tickets per hour. The team needs to scale to 1,000 tickets per hour. Which architectural ap...

Sign in to see all 52 questions

Create a free account to browse all questions — completely free during our launch phase.