Anthropic CCAR-P (Claude Certified Architect - Professional) Practice Test 75

Question 740

The post-incident analysis for software platform concludes that treating drafts and final transmissions as equal risk weakens control. Which remediation targets the relevant failure boundary?

  1. Prioritize Reviewer evidence packet
  2. Prioritize Reversibility-aware approval threshold
  3. Prioritize Risk-tiered approval workflow
  4. Prioritize Random audit of low-risk automation
  5. Prioritize Separation of duties for sensitive changes

Explanation

Correct Answer

B:

The reported failure calls for the control under which the control layer can require stronger approval before externally visible or irreversible steps. A prompt-only reminder or unrelated metric would not address the risk that treating drafts and final transmissions as equal risk weakens control. Evidence of resolution requires explicit side-effect classification and send permissions and an improvement in unauthorized external dispatch rate during staging tests.

Incorrect Answers

A:

This addresses the failure where an approval button without context cannot support an accountable decision, because the control layer can present retrieved evidence, uncertainty and proposed effect alongside each action; but this remedy operates on another boundary, not the case where treating drafts and final transmissions as equal risk weakens control.

C:

When blanket approval for every step wastes capacity; blanket autonomy hides risk, the control layer can send consequential actions to authorized reviewers while allowing bounded automation elsewhere; yet it would not correct the observed weakness: treating drafts and final transmissions as equal risk weakens control.

D:

This addresses the failure where no-review automation can silently degrade after a distribution shift, because the control layer can sample approved outputs for independent periodic quality review; but this remedy operates on another boundary, not the case where treating drafts and final transmissions as equal risk weakens control.

E:

When self-approval eliminates independent control over privileged operations, the control layer can require a distinct authorized human or system role for approval; whereas the software platform defect follows the path in which treating drafts and final transmissions as equal risk weakens control.

Question 741

The defining problem for an internal audit pilot is that short chunks are searchable but answers need full contractual sections. Which proposed approach addresses it most precisely?

  1. Adopt Overlapping chunk strategy
  2. Adopt Semantic section-aware chunking
  3. Adopt Parent-child document retrieval
  4. Adopt Incremental refresh and delete propagation

Explanation

Correct Answer

C:

The mechanism is specific: the integration can search fine-grained chunks then return appropriate parent spans for grounding. A credible implementation requires parent-child mapping and document version IDs, while its acceptance test monitors full-section citation accuracy at a fixed context budget. The alternative approaches leave the risk that returning only tiny isolated fragments removes governing context.

Incorrect Answers

A:

When factual statements often cross page or paragraph cut points, the integration can use measured overlap where splitting otherwise loses neighboring facts; the missing operational behavior is specific: the integration can search fine-grained chunks then return appropriate parent spans for grounding.

B:

When policies contain headings and exceptions that must stay together, the integration can split at meaningful sections and preserve local context around exceptions; the current decision instead depends on a specific capability: the integration can search fine-grained chunks then return appropriate parent spans for grounding.

D:

When source policies change daily and retired versions must stop appearing, the integration can synchronize create-update-delete events into the retrieval index; the missing operational behavior is specific: the integration can search fine-grained chunks then return appropriate parent spans for grounding.

Question 742

A human-resources services incident exposes a specific weakness: fluent restatement can hide intermediate fabrication. Which architectural correction addresses the cause rather than masking the symptom?

  1. Apply Unsupported-claim provenance analysis
  2. Apply Model capability mismatch test
  3. Apply Retrieval recall diagnosis
  4. Apply Tool-result integrity diagnosis

Explanation

Correct Answer

D:

The problem lies in the absence of a mechanism where the evaluation process can compare generated fields with raw tool result identifiers and values. Reinstating it, backed by immutable tool-result capture and lineage IDs, limits the exposure that fluent restatement can hide intermediate fabrication; the engineering team should track output-to-tool-value mismatch frequency.

Incorrect Answers

A:

When adding more citation markers can disguise hallucinations, the evaluation process can validate claim-to-span entailment rather than just link presence; yet it would not correct the observed weakness: fluent restatement can hide intermediate fabrication.

B:

This addresses the failure where deeper retrieval does not fix inadequate synthesis capability, because the evaluation process can evaluate a more capable tier on the same evidence and instructions; but this remedy operates on another boundary, not the case where fluent restatement can hide intermediate fabrication.

C:

This addresses the failure where rewriting prompts cannot compensate for evidence never retrieved, because the evaluation process can inspect whether relevant source passages reached the model; whereas the human-resources services defect follows the path in which fluent restatement can hide intermediate fabrication.

Question 743

The public-sector casework architect proposes omitting Relevance-ranked context packing from the production design. What is the most direct failure mode that decision would expose?

  1. An exposure in which first-in-first-out truncation may discard decisive exceptions
  2. An exposure in which request totals alone cannot isolate token-heavy work
  3. An exposure in which blindly filling the context window can bury critical evidence
  4. An exposure in which raw replay increases cost and may reintroduce superseded requests

Explanation

Correct Answer

C:

Removing the safeguard exposes a failure in which blindly filling the context window can bury critical evidence. The architecture normally counters this because the model configuration can rank excerpts by task relevance and retain traceable source references. A compensating design would require retrieval relevance scoring and citations and evidence from answer support recall under a fixed token budget.

Incorrect Answers

A:

The competing hazard is that first-in-first-out truncation may discard decisive exceptions in a workload where a long research input must retain governing rules and current evidence; but this is not the immediate loss associated with omitting Relevance-ranked context packing: blindly filling the context window can bury critical evidence.

B:

The competing hazard is that request totals alone cannot isolate token-heavy work in a workload where large context traffic causes budget spikes without obvious request-count increases; however, the named safeguard addresses the distinct hazard that blindly filling the context window can bury critical evidence.

D:

The other exposure arises when a long case management dialogue continually replays stale turns: raw replay increases cost and may reintroduce superseded requests. Omitting the requested control instead leaves the more immediate risk that blindly filling the context window can bury critical evidence; but this is not the immediate loss associated with omitting Relevance-ranked context packing: blindly filling the context window can bury critical evidence.

Question 744

After installing Centralized permission broker for agents in a field-service coordination workflow, the team must validate whether it has reduced the exposure that transferring an agent identity must not silently inherit broader privileges. Which measurement is most relevant?

  1. Track privileged tool denials and policy-consistent action logs
  2. Track rubric pass rate after a fixed maximum revision count
  3. Track schema rejection rate and successful artifact joins
  4. Track prerequisite violations and successful handoff rate

Explanation

Correct Answer

A:

An appropriate gate measures privileged tool denials and policy-consistent action logs. It tests the mechanism in which a service-side broker verifies user authority for each proposed action, rather than a cosmetic demonstration. That distinction matters because transferring an agent identity must not silently inherit broader privileges.

Incorrect Answers

B:

When drafted release notes repeatedly miss mandatory restrictions, a separate evaluation pass checks defined criteria and requests capped revisions; but the acceptance gate must instead inspect privileged tool denials and policy-consistent action logs to detect the targeted failure.

C:

Where teams merge agent outputs that refer to inconsistent record identifiers, a relevant check would establish whether common schemas and immutable artifact versions make specialist outputs composable; by contrast, the field-service coordination release needs evidence from privileged tool denials and policy-consistent action logs.

D:

When a reviewer cannot assess a draft before verified citations are available, the architecture can orchestrate a staged dependency graph with explicit prerequisites; however, that proxy would not demonstrate privileged tool denials and policy-consistent action logs for the selected intervention.

Question 745

A contract administration team is advancing Parent-child document retrieval beyond its proof of concept. Engineers expect that the integration can search fine-grained chunks then return appropriate parent spans for grounding. What evidence of setup readiness should be required first?

  1. Confirm source change events and reconciled index versions
  2. Confirm parent-child mapping and document version IDs
  3. Confirm document structure parsing and meaningful boundaries
  4. Confirm normalized metadata fields and ingestion validation
  5. Confirm synchronized lexical and vector indexes

Explanation

Correct Answer

B:

The prerequisite creates the necessary operating boundary: parent-child mapping and document version IDs. The system is intended to work because the integration can search fine-grained chunks then return appropriate parent spans for grounding. A release without that foundation may still suffer returning only tiny isolated fragments removes governing context; use full-section citation accuracy at a fixed context budget to evaluate the result.

Incorrect Answers

A:

Where source policies change daily and retired versions must stop appearing, the integration can synchronize create-update-delete events into the retrieval index; however, the named intervention needs parent-child mapping and document version IDs before its protection can be enforced.

C:

This prerequisite supports a method used when policies contain headings and exceptions that must stay together, because the integration can split at meaningful sections and preserve local context around exceptions; for this release, the team must first verify parent-child mapping and document version IDs.

D:

Where questions require filtering by jurisdiction and policy effective date, the integration can store reliable source metadata for filtering before similarity search; however, the named intervention needs parent-child mapping and document version IDs before its protection can be enforced.

E:

For the separate challenge where users search both exact product codes and paraphrased descriptions, the integration can combine keyword matches with embeddings and rank the merged results; but the production dependency for Parent-child document retrieval is parent-child mapping and document version IDs.

Question 746

A cost-cutting proposal for media licensing eliminates Change-communication plan. Which undesirable consequence would become most credible after that change?

  1. The risk that a single SLA overcommits low-priority work or underprotects critical flows
  2. The risk that anecdotal messages without context are hard to triage
  3. The risk that HTTP availability alone does not measure decision quality
  4. The risk that silent behavior changes can violate expectations and training materials

Explanation

Correct Answer

D:

The relevant failure mode is that silent behavior changes can violate expectations and training materials. The existing design can prevent it when the delivery team can announce consequential changes, effective dates and rollback contacts. The team needs audience list, release calendar and version notes and must monitor issue reports tied to uncommunicated changes to support any claim of equivalent protection.

Incorrect Answers

A:

When internal analytics and customer-facing support have incompatible latency demands, a single SLA overcommits low-priority work or underprotects critical flows; but this is not the immediate loss associated with omitting Change-communication plan: silent behavior changes can violate expectations and training materials.

B:

The competing hazard is that anecdotal messages without context are hard to triage in a workload where users report problems through scattered chats with no reproduction details; however, the named safeguard addresses the distinct hazard that silent behavior changes can violate expectations and training materials.

C:

The other exposure arises when a service meets infrastructure uptime targets while its answers remain unusable: HTTP availability alone does not measure decision quality. Omitting the requested control instead leaves the more immediate risk that silent behavior changes can violate expectations and training materials; whereas removing Change-communication plan primarily exposes the risk that silent behavior changes can violate expectations and training materials.

Question 747

The public-sector casework delivery group must make the following claim executable: the control layer can require stronger approval before externally visible or irreversible steps. Which TWO measures pair the right intervention with the right readiness condition? Choose TWO.

  1. Track severe error escape rate and review workload by tier
  2. Implement Escalation timeout and fail-closed route
  3. Implement Reversibility-aware approval threshold
  4. Establish identity-separated approval roles and tamper-evident logs
  5. Establish explicit side-effect classification and send permissions

Explanation

Correct Answers

C:

The architecture component is justified by its behavior: the control layer can require stronger approval before externally visible or irreversible steps. It targets the exposure that treating drafts and final transmissions as equal risk weakens control, while the complementary selected actions provide readiness and, for a three-part package, evidence.

E:

The prerequisite is explicit side-effect classification and send permissions. It provides the operating preparation for a system in which the control layer can require stronger approval before externally visible or irreversible steps; without this setup, the planned safeguard has not been demonstrated as deployable.

Incorrect Answers

A:

This test would validate a different benefit: the control layer can send consequential actions to authorized reviewers while allowing bounded automation elsewhere; but that indicator does not verify unauthorized external dispatch rate during staging tests for the selected architecture.

B:

This design choice is aimed at cases where approval queues may be unattended overnight; it is effective because the control layer can hold or safely reject high-impact actions until review is complete. The rollout described is instead governed by the case where a workflow can either draft a notice or send it to all customers and requires a different primary intervention; however, the intended intervention requires that the control layer can require stronger approval before externally visible or irreversible steps.

D:

The alternative setup enables a different mechanism: the control layer can require a distinct authorized human or system role for approval; whereas this intervention requires explicit side-effect classification and send permissions as its operating precondition.

Question 748

For an insurance processing controlled rollout, the essential claim is that the evaluation process can compare generated fields with raw tool result identifiers and values. Identify THREE selections that together specify the intervention, the setup dependency and the result to monitor. Choose THREE.

  1. Establish clean holdout examples and example selection log
  2. Implement Model capability mismatch test
  3. Track output-to-tool-value mismatch frequency
  4. Establish immutable tool-result capture and lineage IDs
  5. Implement Tool-result integrity diagnosis
  6. Track unsupported claim rate by generation stage

Explanation

Correct Answers

C:

The selected measure examines output-to-tool-value mismatch frequency at the relevant task boundary. It supplies evidence for the architecture claim that the evaluation process can compare generated fields with raw tool result identifiers and values, while exposing the scenario where fluent restatement can hide intermediate fabrication.

D:

The prerequisite is immutable tool-result capture and lineage IDs. It provides the operating preparation for a system in which the evaluation process can compare generated fields with raw tool result identifiers and values; without this setup, the planned safeguard has not been demonstrated as deployable.

E:

Implementing the intervention changes the operational mechanism so that the evaluation process can compare generated fields with raw tool result identifiers and values. The failure to prevent is that fluent restatement can hide intermediate fabrication; configuration and validation remain separate release obligations.

Incorrect Answers

A:

The alternative setup enables a different mechanism: the evaluation process can test whether examples bias the model to surface patterns rather than policy rules; whereas this intervention requires immutable tool-result capture and lineage IDs as its operating precondition.

B:

This design choice is aimed at cases where complex cross-document inferences fail despite accurate retrieved passages; it is effective because the evaluation process can evaluate a more capable tier on the same evidence and instructions; however, the intended intervention requires that the evaluation process can compare generated fields with raw tool result identifiers and values.

F:

This test would validate a different benefit: the evaluation process can validate claim-to-span entailment rather than just link presence; but that indicator does not verify output-to-tool-value mismatch frequency for the selected architecture.

Question 749

A research administration pilot has enabled Parallel code-review assistance. Because one general review prompt may miss specialist concerns, the release board wants evidence of effectiveness, not just a favorable demonstration. Which metric most directly tests the proposed control?

  1. Monitor changes with full issue-to-approval trace
  2. Monitor regression test pass rate on agent-authored changes
  3. Monitor documentation contract mismatch count
  4. Monitor distinct actionable findings validated by maintainers

Explanation

Correct Answer

D:

An appropriate gate measures distinct actionable findings validated by maintainers. It tests the mechanism in which the engineering workflow can run security, maintainability and API compatibility reviews separately, rather than a cosmetic demonstration. That distinction matters because one general review prompt may miss specialist concerns.

Incorrect Answers

A:

The proposed measure would suit a setting where release teams cannot tell why generated modifications were proposed, since the engineering workflow can attach issue links, tests, diff summaries and reviewer approvals to patches; by contrast, the research administration release needs evidence from distinct actionable findings validated by maintainers.

B:

The proposed measure would suit a setting where agents generate plausible code that breaks existing behavior, since the engineering workflow can require narrow changes with executable tests and review before merge; however, that proxy would not demonstrate distinct actionable findings validated by maintainers for the selected intervention.

C:

The proposed measure would suit a setting where developers ask Claude to generate interface explanations from code, since the engineering workflow can verify generated docs against source and running contracts; but the acceptance gate must instead inspect distinct actionable findings validated by maintainers to detect the targeted failure.

Question 750

The telecommunications support deployment uses Parent-child document retrieval, but executive reports currently count only successful model responses. The residual concern is that returning only tiny isolated fragments removes governing context. Which acceptance measure would reveal real progress?

  1. Track full-section citation accuracy at a fixed context budget
  2. Track time-to-index-freshness and stale-answer incidence
  3. Measure precision of retrieval after authorization filtering
  4. Track recall on exact-code and paraphrase query subsets

Explanation

Correct Answer

A:

The acceptance measure focuses on full-section citation accuracy at a fixed context budget at the task or workflow boundary. Model fluency would not establish whether the integration can search fine-grained chunks then return appropriate parent spans for grounding, and it could conceal cases where returning only tiny isolated fragments removes governing context.

Incorrect Answers

B:

When source policies change daily, retired versions must stop appearing and the integration can synchronize create-update-delete events into the retrieval index; but the acceptance gate must instead inspect full-section citation accuracy at a fixed context budget to detect the targeted failure.

C:

When questions require filtering by jurisdiction, policy effective date and the integration can store reliable source metadata for filtering before similarity search; by contrast, the telecommunications support release needs evidence from full-section citation accuracy at a fixed context budget.

D:

Where users search both exact product codes and paraphrased descriptions, a relevant check would establish whether the integration can combine keyword matches with embeddings and rank the merged results; however, that proxy would not demonstrate full-section citation accuracy at a fixed context budget for the selected intervention.

Claude Certified Architect - Professional (CCAR-P) Practice Tests