Question 720
A program architect at public-sector casework has endorsed Relevance-ranked context packing. The proposed technical effect is that the model configuration can rank excerpts by task relevance and retain traceable source references. Which implementation condition is necessary to realize that effect?
- Verify explicit summary refresh and fact provenance
- Verify retrieval relevance scoring and citations
- Verify request-level usage accounting and trace identifiers
- Verify typed memory entries and expiry rules
Explanation
Correct Answer
B:What makes the control operable is retrieval relevance scoring and citations. It supplies the precondition for the mechanism whereby the model configuration can rank excerpts by task relevance and retain traceable source references. The residual failure to guard against is that blindly filling the context window can bury critical evidence, measured through answer support recall under a fixed token budget.
Incorrect Answers
A:Where a long case management dialogue continually replays stale turns, the model configuration can preserve verified decisions and current state while pruning redundant history; for this release, the team must first verify retrieval relevance scoring and citations.
C:The setup is useful to enable the alternative's behavior where the model configuration can record uncached input, cache reads, output and tool-related usage per task and large context traffic causes budget spikes without obvious request-count increases; but the production dependency for Relevance-ranked context packing is retrieval relevance scoring and citations.
D:This prerequisite supports a method used when user preferences and external tool outputs enter the same conversation memory, because the model configuration can store verified state separately from untrusted retrieved text and time-bound hints; for this release, the team must first verify retrieval relevance scoring and citations.
Question 721
A healthcare operations incident exposes a specific weakness: unbounded self-critique loops can raise cost without a convergence rule. Which architectural correction addresses the cause rather than masking the symptom?
- Introduce Centralized permission broker for agents
- Introduce Budgeted orchestration stop policy
- Introduce Independent critic with bounded revisions
- Introduce Parallel independent subtask fan-out
Explanation
Correct Answer
C:The remediation acts at the relevant design boundary: a separate evaluation pass checks defined criteria and requests capped revisions. That counteracts the observed hazard that unbounded self-critique loops can raise cost without a convergence rule. Rollout should include an independent rubric and revision limit, with regression evidence from rubric pass rate after a fixed maximum revision count.
Incorrect Answers
A:This addresses the failure where transferring an agent identity must not silently inherit broader privileges, because a service-side broker verifies user authority for each proposed action; yet it would not correct the observed weakness: unbounded self-critique loops can raise cost without a convergence rule.
B:When a task can exhaust resources when delegation has no bounded horizon, the architecture can impose step, token, cost and time ceilings with explicit exit behavior; but this remedy operates on another boundary, not the case where unbounded self-critique loops can raise cost without a convergence rule.
D:Although the alternative reduces the risk that serial execution creates avoidable critical-path latency because the architecture can run independent checks concurrently and combine results with deterministic joins; yet it would not correct the observed weakness: unbounded self-critique loops can raise cost without a convergence rule.
Question 722
A supply-chain planning technical leadership group is designing a Claude-supported workflow because multiple quality concerns need independent assessment of a large change. Which architectural decision most directly meets that requirement?
- Bounded issue-to-patch automation
- Reviewed automated code documentation
- Parallel code-review assistance
- Change provenance for agent edits
Explanation
Correct Answer
C:The intervention acts on the system behavior directly: the engineering workflow can run security, maintainability and API compatibility reviews separately. Before rollout, verify fixed review rubrics and deduplicated findings. The relevant success test is distinct actionable findings validated by maintainers; otherwise one general review prompt may miss specialist concerns.
Incorrect Answers
A:When routine low-risk bugs have reproducible fixtures, the engineering workflow can allow an agent to propose a patch and tests within a restricted repository scope; whereas the supply-chain planning case must guard against the outcome that one general review prompt may miss specialist concerns.
B:If developers ask Claude to generate interface explanations from code, the engineering workflow can verify generated docs against source and running contracts; this rollout still needs the following guarantee: the engineering workflow can run security, maintainability and API compatibility reviews separately.
D:For a workload where release teams cannot tell why generated modifications were proposed, the engineering workflow can attach issue links, tests, diff summaries and reviewer approvals to patches; whereas the supply-chain planning case must guard against the outcome that one general review prompt may miss specialist concerns.
Question 723
A customer success team is selecting the next improvement for its AI service; an agent selects correct tools but invents values absent from responses. What design choice has the strongest direct justification?
- Tool-result integrity diagnosis
- Model capability mismatch test
- Overfit-to-examples analysis
- Context truncation investigation
- Unsupported-claim provenance analysis
Explanation
Correct Answer
A:The architectural distinction is that the evaluation process can compare generated fields with raw tool result identifiers and values. The decision should be implemented with immutable tool-result capture and lineage IDs and evaluated against output-to-tool-value mismatch frequency; otherwise, fluent restatement can hide intermediate fabrication.
Incorrect Answers
B:If complex cross-document inferences fail despite accurate retrieved passages, the evaluation process can evaluate a more capable tier on the same evidence and instructions; it does not ensure the required behavior: the evaluation process can compare generated fields with raw tool result identifiers and values.
C:If a few-shot classifier works on familiar categories but fails novel wording, the evaluation process can test whether examples bias the model to surface patterns rather than policy rules; whereas the media licensing case must guard against the outcome that fluent restatement can hide intermediate fabrication.
D:If failures spike only when lengthy contracts include appendices, the evaluation process can inspect token budgets and the presence of decisive passages after packing; the alternative leaves the central requirement unproven: the evaluation process can compare generated fields with raw tool result identifiers and values.
E:For a workload where a response confidently cites text that does not support the claim, the evaluation process can validate claim-to-span entailment rather than just link presence; it does not ensure the required behavior: the evaluation process can compare generated fields with raw tool result identifiers and values.
Question 724
A logistics coordination service owner is reviewing Incremental refresh and delete propagation after a canary rollout. The critical regression would occur if stale embeddings can support obsolete guidance after source correction. What observable acceptance signal belongs in the dashboard?
- Track time-to-index-freshness and stale-answer incidence
- Track full-section citation accuracy at a fixed context budget
- Track recall on exact-code and paraphrase query subsets
- Track eligible-document retrieval precision after filtering
Explanation
Correct Answer
A:Tracking time-to-index-freshness and stale-answer incidence provides direct evidence that the system delivers its intended effect: the integration can synchronize create-update-delete events into the retrieval index. The measure is more informative than prompt counts because the failure being measured is that stale embeddings can support obsolete guidance after source correction.
Incorrect Answers
B:Where short chunks are searchable but answers need full contractual sections, a relevant check would establish whether the integration can search fine-grained chunks then return appropriate parent spans for grounding; by contrast, the logistics coordination release needs evidence from time-to-index-freshness and stale-answer incidence.
C: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 time-to-index-freshness and stale-answer incidence for the selected intervention.
D:Where questions require filtering by jurisdiction and policy effective date, a relevant check would establish whether the integration can store reliable source metadata for filtering before similarity search; but the acceptance gate must instead inspect time-to-index-freshness and stale-answer incidence to detect the targeted failure.
Question 725
For an energy operations rollout, the proposed design uses Change-communication plan, relying on the claim that the delivery team can announce consequential changes, effective dates and rollback contacts. What enabling control must be in place to make that design enforceable?
- Confirm common budget definition and alert thresholds
- Confirm ground-truth validation and service owner
- Confirm on-call roster and incident criteria
- Confirm audience list, release calendar and version notes
Explanation
Correct Answer
D:The proposed design needs audience list, release calendar and version notes as its foundation. This matters because the delivery team can announce consequential changes, effective dates and rollback contacts; a plausible demonstration without the prerequisite can still leave a case where silent behavior changes can violate expectations and training materials. The relevant check is issue reports tied to uncommunicated changes.
Incorrect Answers
A:For the separate challenge where a deployment repeatedly misses agreed quality targets, the delivery team can review error budget consumption and pause risky launches when exhausted; but the production dependency for Change-communication plan is audience list, release calendar and version notes.
B:The setup is useful to enable the alternative's behavior where the delivery team can specify user-observable success such as validated resolutions and response time and a service meets infrastructure uptime targets while its answers remain unusable; for this release, the team must first verify audience list, release calendar and version notes.
C:For the separate challenge where users do not know whom to call when generated advice causes harm, the delivery team can map severity levels to operational, security and business response roles; however, the named intervention needs audience list, release calendar and version notes before its protection can be enforced.
Question 726
The solution design for enterprise knowledge management assumes that the control layer can require stronger approval before externally visible or irreversible steps. Before treating Reversibility-aware approval threshold as production-ready, which dependency needs an owner and verification?
- Confirm concise reviewer reason codes and privacy controls
- Confirm identity-separated approval roles and tamper-evident logs
- Confirm explicit side-effect classification and send permissions
- Confirm source traceability and readable audit interface
Explanation
Correct Answer
C:Readiness depends on explicit side-effect classification and send permissions, not just a design approval. That condition supports the mechanism in which the control layer can require stronger approval before externally visible or irreversible steps. Without it, treating drafts and final transmissions as equal risk weakens control; monitoring unauthorized external dispatch rate during staging tests helps verify the deployed result.
Incorrect Answers
A:For the separate challenge where reviewers often correct suggestions but no explanation is recorded, the control layer can capture structured reasons and attach them to evaluation feedback; for this release, the team must first verify explicit side-effect classification and send permissions.
B:For the separate challenge where the same assistant could propose and approve high-impact access changes, the control layer can require a distinct authorized human or system role for approval; however, the named intervention needs explicit side-effect classification and send permissions before its protection can be enforced.
D:The setup is useful to enable the alternative's behavior where the control layer can present retrieved evidence, uncertainty and proposed effect alongside each action and approvers receive model recommendations without underlying sources; for this release, the team must first verify explicit side-effect classification and send permissions.
Question 727
A field-service coordination rollout must satisfy this design premise: a separate evaluation pass checks defined criteria and requests capped revisions. Choose TWO linked items that form a workable architecture change and its essential preparation. Choose TWO.
- Track privileged tool denials and policy-consistent action logs
- Establish an independent rubric and revision limit
- Implement Independent critic with bounded revisions
- Implement Sequential dependency-aware handoffs
- Establish role-specific tool grants and aggregation rules
Explanation
Correct Answers
B:This step creates the enabling basis of an independent rubric and revision limit. That preparation supports the mechanism that a separate evaluation pass checks defined criteria and requests capped revisions; merely choosing the control does not prove its configuration or authority is ready.
C:The architecture component is justified by its behavior: a separate evaluation pass checks defined criteria and requests capped revisions. It targets the exposure that unbounded self-critique loops can raise cost without a convergence rule, while the complementary selected actions provide readiness and, for a three-part package, evidence.
Incorrect Answers
A:This test would validate a different benefit: a service-side broker verifies user authority for each proposed action; but that indicator does not verify rubric pass rate after a fixed maximum revision count for the selected architecture.
D:This design choice is aimed at cases where a reviewer cannot assess a draft before verified citations are available; however, the intended intervention requires that a separate evaluation pass checks defined criteria and requests capped revisions.
E:The alternative setup enables a different mechanism: a supervisor assigns tasks and consolidates results within distinct permission scopes; whereas this intervention requires an independent rubric and revision limit as its operating precondition.
Question 728
A healthcare operations pilot has enabled Relevance-ranked context packing. Because blindly filling the context window can bury critical evidence, the release board wants evidence of effectiveness, not just a favorable demonstration. Which metric most directly tests the proposed control?
- Audit context-overflow rate at high-percentile request size
- Audit tokens and dollars per accepted outcome by category
- Audit answer support recall under a fixed token budget
- Audit retained authoritative facts after context trimming
Explanation
Correct Answer
C:The acceptance measure focuses on answer support recall under a fixed token budget at the task or workflow boundary. Model fluency would not establish whether the model configuration can rank excerpts by task relevance and retain traceable source references, and it could conceal cases where blindly filling the context window can bury critical evidence.
Incorrect Answers
A:Where templates unexpectedly exceed request limits after tool results are appended, a relevant check would establish whether the model configuration can estimate input, tool-result and output allowances before sending requests; by contrast, the healthcare operations release needs evidence from answer support recall under a fixed token budget.
B:Where large context traffic causes budget spikes without obvious request-count increases, a relevant check would establish whether the model configuration can record uncached input, cache reads, output and tool-related usage per task; however, that proxy would not demonstrate answer support recall under a fixed token budget for the selected intervention.
D:When a long research input must retain governing rules, current evidence and the model configuration can truncate lowest-priority material while protecting essential instructions and citations. A favorable result on that test would not establish answer support recall under a fixed token budget for this workflow, where a case includes more background than the model can productively use; by contrast, the healthcare operations release needs evidence from answer support recall under a fixed token budget.
Question 729
A field-service coordination implementation team has approved Tool-result integrity diagnosis. Its expected behavior is that the evaluation process can compare generated fields with raw tool result identifiers and values. What prerequisite belongs on the release-readiness checklist?
- Verify immutable tool-result capture and lineage IDs
- Verify serialized prompt capture with redaction
- Verify source spans and expert claim verification
- Verify clean holdout examples and example selection log
Explanation
Correct Answer
A:Before deployment the team must confirm immutable tool-result capture and lineage IDs. Without that preparation, the claim that the evaluation process can compare generated fields with raw tool result identifiers and values may remain unenforceable; the related exposure is that fluent restatement can hide intermediate fabrication. The acceptance indicator is output-to-tool-value mismatch frequency.
Incorrect Answers
B:For the separate challenge where failures spike only when lengthy contracts include appendices, the evaluation process can inspect token budgets and the presence of decisive passages after packing; however, the named intervention needs immutable tool-result capture and lineage IDs before its protection can be enforced.
C:The setup is useful to enable the alternative's behavior where the evaluation process can validate claim-to-span entailment rather than just link presence and a response confidently cites text that does not support the claim; but the production dependency for Tool-result integrity diagnosis is immutable tool-result capture and lineage IDs.
D:Where a few-shot classifier works on familiar categories but fails novel wording, the evaluation process can test whether examples bias the model to surface patterns rather than policy rules; for this release, the team must first verify immutable tool-result capture and lineage IDs.
Question 730
An independent assessment of a logistics coordination Claude system reports that stale embeddings can support obsolete guidance after source correction. Which change best closes the exposed control gap?
- Deploy Incremental refresh and delete propagation
- Deploy Semantic section-aware chunking
- Deploy Overlapping chunk strategy
- Deploy Metadata-rich index entries
Explanation
Correct Answer
A:This is the causal correction: the integration can synchronize create-update-delete events into the retrieval index. With source change events and reconciled index versions in place, the team can test time-to-index-freshness and stale-answer incidence and show that the problematic case in which stale embeddings can support obsolete guidance after source correction has been addressed.
Incorrect Answers
B:When fixed arbitrary cuts can detach qualifiers from the associated rule, the integration can split at meaningful sections and preserve local context around exceptions; but this remedy operates on another boundary, not the case where stale embeddings can support obsolete guidance after source correction.
C:This addresses the failure where excessive overlap inflates storage and duplicates retrieved context, because the integration can use measured overlap where splitting otherwise loses neighboring facts; whereas the logistics coordination defect follows the path in which stale embeddings can support obsolete guidance after source correction.
D:Where questions require filtering by jurisdiction and policy effective date, the integration can store reliable source metadata for filtering before similarity search; yet it would not correct the observed weakness: stale embeddings can support obsolete guidance after source correction.