CCAR-F practice › Agentic Architecture & Orchestration · Question 16 of 105
Agentic Architecture & Orchestration · Select TWO
A support agent's system prompt contains the four rules below. Which two should be enforced programmatically rather than left to prompt guidance? Select two.
Show answer and explanation
Answer: A and C — Verify the customer's identity with get_customer before calling process_refund / Never issue a refund above the agent's approval limit without escalation
Rules that protect identity and money require deterministic compliance, which only programmatic enforcement (prerequisite gates, interception hooks) provides; prompt instructions have a non-zero failure rate. Tone (b) and the order in which to offer acceptable options (d) are preferences where occasional deviation is tolerable, so prompt guidance is appropriate.
Keep going
More Agentic Architecture & Orchestration questions
- Where should enforcement of a rule like "never transfer more than $10,000 without a…
- A support agent pulls order data from an MCP tool that returns Unix timestamps and…
- Policy requires human approval for refunds over $500. A team adds a hook that rejects…
- An agent team is deciding where to use Agent SDK hooks instead of prompt instructions…
- A CI step reviews pull requests by sending all changed files to Claude in one prompt. On…