CCAR-F Agentic Architecture and Orchestration Practice Question
An architect is configuring a Claude agent that must decide when to search an internal knowledge base versus when to answer directly from its own knowledge. The team wants the agent to retrieve only when the question references internal policies, products, or procedures. Which configuration most directly supports this decision?
⚠ Common exam trap
The trap here is conflating tool_choice any with auto; forcing any guarantees retrieval on every turn, which is the opposite of selective retrieval.
Answer choices
Why each option matters
Answer the question above first, then reveal the full breakdown to understand why each option is right or wrong.
Correct answer & explanation
✓
Set tool_choice to auto and provide a well-described search tool so Claude decides per turn whether retrieval is needed.
Selective retrieval is achieved by letting the model decide per turn using tool_choice auto plus a precise tool description. The description tells Claude what content the tool covers, so it invokes search for internal policy, product, or procedure questions and answers directly otherwise. Forcing tool use, disabling tools, or gating on a keyword all remove that contextual judgment and produce either over-retrieval or missed retrieval.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Register the search tool but instruct Claude to call it only when the user explicitly types the word search.
Why it's wrong here
Tying retrieval to a literal keyword makes the agent brittle and dependent on user phrasing. Users rarely know when internal knowledge is required, which is precisely the judgment the agent should make. This rule also prevents retrieval for legitimate policy questions that omit the keyword, defeating the purpose of an autonomous retrieval decision.
- ✓
Set tool_choice to auto and provide a well-described search tool so Claude decides per turn whether retrieval is needed.
Why this is correct
With tool_choice set to auto, Claude evaluates the user's request and the tool description each turn and emits a tool_use block only when the description matches the need. A clear description explaining that the tool covers internal policies, products, and procedures gives the model the signal it needs to retrieve selectively rather than on every turn.
- ✗
Disable tools entirely and rely on a large system prompt containing the most frequently asked policy excerpts.
Why it's wrong here
Embedding excerpts in the system prompt cannot cover a dynamic or large knowledge base and will go stale as policies change. It also consumes context on every request and offers no mechanism to fetch content the prompt omits. Removing tools eliminates the agent's ability to retrieve current internal information, directly undermining the stated goal.
- ✗
Set tool_choice to any and register only the knowledge base search tool so the model must always retrieve before answering.
Why it's wrong here
Forcing tool use makes every turn retrieve, including questions the model could answer directly from general knowledge. This contradicts the requirement to retrieve only for internal references. It also adds latency and cost for simple queries and can inject irrelevant context that degrades answer quality, so it does not match the desired selective behavior.
About these practice questions
Courseiva writes every CCAR-F question from scratch — 271 in total, each with an explanation and a wrong-answer breakdown. None are copied from real exams or dumps. Learn why practice questions differ from exam dumps →
JA
Written and reviewed by Johnson Ajibi, MSc IT Security
Senior Network & Security Engineer · founder of Courseiva
Last reviewed September 2026 · checked against the official Anthropic exam blueprint
This CCAR-F practice question is part of Courseiva's free Anthropic certification practice question bank. Courseiva provides original exam-style practice questions with explanations, topic-based practice, mock exams, readiness tracking, and study analytics to help learners prepare for the CCAR-F exam.