CCAR-F Context and Reliability Practice Question
A travel assistant maintains a long conversation with a user over several days. The system prompt defines a strict booking policy, but by the fortieth turn Claude begins offering refunds that violate that policy. The team wants to keep the full history for continuity while preventing policy erosion. Which architecture best addresses this?
⚠ Common exam trap
The trap here is assuming that a longer conversation needs more memory or output space, when the real cause is that the governing instruction loses salience as the transcript grows.
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
✓
Re-assert the booking policy as a system-level instruction periodically or before policy-sensitive actions.
Policy drift over long conversations is countered by keeping the governing instruction salient near the decision point. Periodically re-asserting the booking policy, or injecting it before policy-sensitive actions, preserves full history while restoring the constraint's influence. Truncation, larger output limits, and user-driven reminders do not address the loss of instruction salience across many turns.
Answer analysis
Option-by-option breakdown
For each option: why learners choose it and why it is or isn't the right answer here.
- ✗
Delete older turns once the conversation exceeds thirty messages so only recent context remains.
Why it's wrong here
Truncating history discards the continuity the team explicitly wants to preserve and can drop details the user expects Claude to remember. It also does not reinforce the policy, which is the actual failure. The problem is that the governing instruction loses influence over a long transcript, not that the transcript is too long. Aggressive deletion trades one defect for another without fixing policy adherence.
- ✓
Re-assert the booking policy as a system-level instruction periodically or before policy-sensitive actions.
Why this is correct
Restating the governing policy near the point of decision keeps it salient as the conversation grows, counteracting the drift that occurs when instructions sit far from the current turn. This preserves full history for continuity while reinforcing the constraint exactly when it matters. It directly targets the observed erosion of policy adherence across many turns without discarding context the user relies on.
- ✗
Ask the user to restate the booking policy at the start of each session.
Why it's wrong here
Shifting responsibility to the user is unreliable and poor design; users should not have to enforce the system's own rules. It also does not guarantee correct behavior mid-conversation where the drift occurs. The architecture should enforce policy independently of user cooperation. This approach adds friction and still leaves the underlying salience problem unresolved.
- ✗
Increase the model's max_tokens so it has more room to recall the policy when answering.
Why it's wrong here
Max tokens limits output length and does not expand how well Claude retains or weights earlier instructions. The policy is already present in the system prompt; the issue is its diminishing influence over a long transcript. Granting more room for the reply cannot restore adherence to a constraint the model is underweighting. This change addresses output size, not instruction salience.
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.