Question 370
A program architect at manufacturing quality has endorsed Automation bias countermeasure. The proposed technical effect is that the control layer can design independent checks and reason-coded overrides for high-stakes cases. Which implementation condition is necessary to realize that effect?
- Arrange uncertainty rubric and explanation requirements
- Arrange gold answers and claim-verification criteria
- Arrange approved data-flow inventory and service agreements
- Arrange training, reasonable workload and escalation authority
Explanation
Correct Answer
D:The required preparation is training, reasonable workload and escalation authority. It makes the following architecture claim enforceable: the control layer can design independent checks and reason-coded overrides for high-stakes cases. If that setup is skipped, a human in the workflow may not provide meaningful oversight, so the release should examine reviewer disagreement and caught-error rates in seeded tests.
Incorrect Answers
A:Where an assistant states confident conclusions without revealing evidence limitations, the control layer can require source grounding and explicit uncertainty when proof is unavailable; however, the named intervention needs training, reasonable workload and escalation authority before its protection can be enforced.
B:Where generated citations may be fluent yet unsupported by source passages, the control layer can assess error severity and enforce evidence or abstention at decision points; but the production dependency for Automation bias countermeasure is training, reasonable workload and escalation authority.
C:Where customer records are sent to external inference and telemetry systems, the control layer can trace data paths, access, retention and deletion obligations before launch; for this release, the team must first verify training, reasonable workload and escalation authority.
Question 371
A manufacturing quality incident exposes a specific weakness: holding an interactive connection open invites timeouts and duplicate retries. Which architectural correction addresses the cause rather than masking the symptom?
- Apply Asynchronous completion for long tasks
- Apply Selective verification for risky outputs
- Apply Latency-aware fallback answer
- Apply Parallel independent retrieval
- Apply Rerank only the candidate set
Explanation
Correct Answer
A:The engineering fix changes the mechanism rather than the presentation: the integration can submit a job and return an identifier with status tracking. It requires persistent job state and idempotent submission; a regression test should examine completion success and queue wait under load in cases where holding an interactive connection open invites timeouts and duplicate retries.
Incorrect Answers
B:The alternative reduces the risk that uniform verification can miss the interactive latency objective because the integration can apply expensive verification only to high-impact or low-confidence actions; however, the manufacturing quality defect follows the path in which holding an interactive connection open invites timeouts and duplicate retries.
C:When cutting retrieval silently may produce unsupported definitive statements, the integration can provide an explicit partial or queued result rather than fabricate a quick full answer; yet it would not correct the observed weakness: holding an interactive connection open invites timeouts and duplicate retries.
D:Where an answer needs evidence from several unrelated repositories, the integration can fetch independent sources concurrently and join their authorized results; but this remedy operates on another boundary, not the case where holding an interactive connection open invites timeouts and duplicate retries.
E:Where broad retrieval finds relevant content but top results contain noisy matches, the integration can rerank a bounded candidate list for precision before generation; whereas the manufacturing quality defect follows the path in which holding an interactive connection open invites timeouts and duplicate retries.
Question 372
The software platform staging review finds that static snapshots fail to reveal stale-index behavior. Which design intervention is the most appropriate corrective action?
- Apply Stratified production-like test set
- Apply Metamorphic consistency tests
- Apply Temporal holdout for changing policies
- Apply Held-out evaluation split
- Apply Adversarial injection fixtures
Explanation
Correct Answer
C:The reported failure calls for the control under which the evaluation process can test on later source changes to measure freshness and drift response. A prompt-only reminder or unrelated metric would not address the risk that static snapshots fail to reveal stale-index behavior. Evidence of resolution requires dated corpora and known change events and an improvement in version-correct answers after policy refresh.
Incorrect Answers
A:This addresses the failure where a high average score can mask catastrophic minority failures, because the evaluation process can stratify by task type, risk, language and source complexity; whereas the software platform defect follows the path in which static snapshots fail to reveal stale-index behavior.
B:Although the alternative reduces the risk that a single fixed example cannot expose brittle prompt sensitivity because the evaluation process can apply controlled input transformations and compare invariant outcomes; yet it would not correct the observed weakness: static snapshots fail to reveal stale-index behavior.
D:The alternative reduces the risk that reused training examples inflate measured generalization because the evaluation process can reserve untouched cases for final comparison after iteration; however, the software platform defect follows the path in which static snapshots fail to reveal stale-index behavior.
E:When normal benign examples do not exercise trust boundaries, the evaluation process can add attacker-controlled content and escalation attempts to evaluation; yet it would not correct the observed weakness: static snapshots fail to reveal stale-index behavior.
Question 373
A lean redesign in logistics coordination would retire Guardrail outside generation while leaving the rest of the service unchanged. What consequence makes this change unsafe without replacement controls?
- The risk that prose instructions alone can permit inconsistent formatting
- The risk that a warning in the prompt does not prevent a tool executor from making a prohibited change
- The risk that promoting document content into authority enables prompt injection
- The risk that an untracked hot edit invalidates comparison with prior test results
Explanation
Correct Answer
B:The risk is not a generic model inconvenience: a warning in the prompt does not prevent a tool executor from making a prohibited change. The proposed safeguard changes that boundary because the model configuration can enforce authorization and policy in deterministic application controls, and the release would need server-side policy enforcement before side effects plus prevented unauthorized actions during adversarial tests to show the risk remains contained.
Incorrect Answers
A:The other exposure arises when downstream clients need stable currency: date conventions and prose instructions alone can permit inconsistent formatting. Omitting the requested control instead leaves the more immediate risk that a warning in the prompt does not prevent a tool executor from making a prohibited change; whereas removing Guardrail outside generation primarily exposes the risk that a warning in the prompt does not prevent a tool executor from making a prohibited change.
C:When retrieved customer text contains hostile instructions to override policy, promoting document content into authority enables prompt injection; however, the named safeguard addresses the distinct hazard that a warning in the prompt does not prevent a tool executor from making a prohibited change.
D:When deployments silently change instruction wording between evaluation and production, an untracked hot edit invalidates comparison with prior test results; whereas removing Guardrail outside generation primarily exposes the risk that a warning in the prompt does not prevent a tool executor from making a prohibited change.
Question 374
Following a regression in human-resources services, the team establishes that a network timeout can trigger repeated external state changes. Which proposed redesign most directly prevents recurrence?
- Introduce Idempotent action boundary
- Introduce Typed output validation and repair route
- Introduce Preprocessing and sensitive-field minimization
- Introduce Feedback taxonomy with ground-truth capture
Explanation
Correct Answer
A:The remediation acts at the relevant design boundary: idempotency keys and deduplication protect side-effecting writes across retries. That counteracts the observed hazard that a network timeout can trigger repeated external state changes. Rollout should include stable request identifiers and a deduplication store, with regression evidence from duplicate external writes per retried request.
Incorrect Answers
B:The alternative reduces the risk that accepting plausible-looking prose can break consumers or misroute cases because the architecture can validate generated payloads against a contract and route failures to repair or review; however, the human-resources services defect follows the path in which a network timeout can trigger repeated external state changes.
C:This addresses the failure where sending raw records expands exposure without helping classification, because preprocessing redacts unnecessary fields while preserving task-bearing content; yet it would not correct the observed weakness: a network timeout can trigger repeated external state changes.
D:This addresses the failure where undifferentiated feedback cannot guide targeted changes, because structured error tags make future fixes attributable to retrieval, prompt or policy causes; but this remedy operates on another boundary, not the case where a network timeout can trigger repeated external state changes.
Question 375
The software platform release team needs a numerical gate for Audience-specific architecture views. Rather than equating deployment with success, it wants to catch cases in which a single highly technical diagram can obscure business consequences. Which indicator should it track?
- Audit decision quality against agreed outage and peak-load scenarios
- Audit scenario-adjusted total ownership cost over the planning horizon
- Audit cross-audience agreement on boundaries and commitments
- Audit decisions traced to approved evidence and reversal triggers
Explanation
Correct Answer
C:An appropriate gate measures cross-audience agreement on boundaries and commitments. It tests the mechanism in which the delivery team can use one consistent design with distinct decision and implementation views, rather than a cosmetic demonstration. That distinction matters because a single highly technical diagram can obscure business consequences.
Incorrect Answers
A:Where two proposed architectures differ in recovery behavior during connector outages, a relevant check would establish whether the delivery team can compare alternatives using concrete failure scenarios rather than vague preferences; however, that proxy would not demonstrate cross-audience agreement on boundaries and commitments for the selected intervention.
B:The alternative indicator checks whether the delivery team can include maintenance, migration, vendor dependencies and personnel effort in comparison when a model switch looks cheap but requires extensive template and integration rework; but the acceptance gate must instead inspect cross-audience agreement on boundaries and commitments to detect the targeted failure.
D:The alternative indicator checks whether the delivery team can record context, alternatives, decision and consequences in a versioned ADR when teams repeatedly reopen the same model-versus-workflow decision; however, that proxy would not demonstrate cross-audience agreement on boundaries and commitments for the selected intervention.
Question 376
During an acceptance review in logistics coordination, Automation bias countermeasure is operational but its benefit has not been measured. Auditors are concerned that a human in the workflow may not provide meaningful oversight. Which evidence most directly settles the question?
- Evaluate unsupported claims weighted by business severity
- Evaluate reviewer disagreement and caught-error rates in seeded tests
- Evaluate policy regression detection time after controlled changes
- Evaluate unauthorized cross-service propagation attempts blocked
- Evaluate successful cross-boundary actions on red-team scenarios
Explanation
Correct Answer
B:The acceptance measure focuses on reviewer disagreement and caught-error rates in seeded tests at the task or workflow boundary. Model fluency would not establish whether the control layer can design independent checks and reason-coded overrides for high-stakes cases, and it could conceal cases where a human in the workflow may not provide meaningful oversight.
Incorrect Answers
A:The alternative indicator checks whether the control layer can assess error severity and enforce evidence or abstention at decision points when generated citations may be fluent yet unsupported by source passages; but the acceptance gate must instead inspect reviewer disagreement and caught-error rates in seeded tests to detect the targeted failure.
C:Where new prompt versions or changing input mix alter policy compliance, a relevant check would establish whether the control layer can establish post-release monitoring and re-evaluation triggers; however, that proxy would not demonstrate reviewer disagreement and caught-error rates in seeded tests for the selected intervention.
D:Where multiple connectors relay credentials and untrusted results, a relevant check would establish whether the control layer can identify propagation paths for compromised tokens and manipulated outputs; but the acceptance gate must instead inspect reviewer disagreement and caught-error rates in seeded tests to detect the targeted failure.
E:Where an agent reads content written by unknown third parties, a relevant check would establish whether the control layer can enumerate trust-boundary crossings and test tool redirection attempts; by contrast, the logistics coordination release needs evidence from reviewer disagreement and caught-error rates in seeded tests.
Question 377
A research administration service passed its happy-path demo, and a reviewer asks to discard Asynchronous completion for long tasks. Which concern should the approval board raise?
- The danger that uniform verification can miss the interactive latency objective
- The danger that a fixed high budget can waste time and tokens on simple cases
- The danger that sequential calls inflate end-to-end wait time without improving correctness
- The danger that holding an interactive connection open invites timeouts and duplicate retries
Explanation
Correct Answer
D:The relevant failure mode is that holding an interactive connection open invites timeouts and duplicate retries. The existing design can prevent it when the integration can submit a job and return an identifier with status tracking. The team needs persistent job state and idempotent submission and must monitor completion success and queue wait under load to support any claim of equivalent protection.
Incorrect Answers
A:If routine descriptions are low stakes but payment changes require evidence checking, uniform verification can miss the interactive latency objective; but this is not the immediate loss associated with omitting Asynchronous completion for long tasks: holding an interactive connection open invites timeouts and duplicate retries.
B:The other exposure arises when straightforward lookup requests: deep analysis use very different effort and a fixed high budget can waste time and tokens on simple cases; however, the named safeguard addresses the distinct hazard that holding an interactive connection open invites timeouts and duplicate retries.
C:The other exposure arises when an answer needs evidence from several unrelated repositories: sequential calls inflate end-to-end wait time without improving correctness; whereas removing Asynchronous completion for long tasks primarily exposes the risk that holding an interactive connection open invites timeouts and duplicate retries.
Question 378
A procurement services rollout must satisfy this design premise: the evaluation process can test on later source changes to measure freshness and drift response. Choose TWO linked items that form a workable architecture change and its essential preparation. Choose TWO.
- Establish dated corpora and known change events
- Establish leakage-resistant dataset versioning
- Track invariance violations across controlled transformations
- Implement Stratified production-like test set
- Implement Temporal holdout for changing policies
Explanation
Correct Answers
A:Choosing this setup item establishes dated corpora and known change events, which is needed for reliable operation when the evaluation process can test on later source changes to measure freshness and drift response. The Temporal holdout for changing policies rollout should not be declared ready until dated corpora and known change events is verified in its operating environment.
E:The control itself makes the following capability possible: the evaluation process can test on later source changes to measure freshness and drift response. That matters when static snapshots fail to reveal stale-index behavior; it does not replace the required enabling condition or outcome test.
Incorrect Answers
B:The alternative setup enables a different mechanism: the evaluation process can reserve untouched cases for final comparison after iteration; whereas this intervention requires dated corpora and known change events as its operating precondition.
C:This test would validate a different benefit: the evaluation process can apply controlled input transformations and compare invariant outcomes; but that indicator does not verify version-correct answers after policy refresh for the selected architecture.
D:This design choice is aimed at cases where long-tail failure classes rarely appear in a random convenience sample; it is effective because the evaluation process can stratify by task type, risk, language and source complexity. The rollout described is instead governed by the case where policy retrieval quality looks good using documents from the test period and requires a different primary intervention; however, the intended intervention requires that the evaluation process can test on later source changes to measure freshness and drift response.
Question 379
A manufacturing quality implementation team has approved Workspace separation for sensitive tasks. Its expected behavior is that the engineering workflow can isolate project roots, secrets and tool permissions by approved workspace. What prerequisite belongs on the release-readiness checklist?
- Verify central connector catalog and credential lifecycle
- Verify approved project guidance and code review owners
- Verify isolated identities and explicit workspace inventory
- Verify authorized hooks and maintained check commands
Explanation
Correct Answer
C:What makes the control operable is isolated identities and explicit workspace inventory. It supplies the precondition for the mechanism whereby the engineering workflow can isolate project roots, secrets and tool permissions by approved workspace. The residual failure to guard against is that a shared default environment increases cross-project exposure, measured through cross-workspace access violations in test cases.
Incorrect Answers
A:Where development teams connect tools with overlapping credentials and contracts, the engineering workflow can register reviewed MCP servers with clear ownership and allowed scopes; however, the named intervention needs isolated identities and explicit workspace inventory before its protection can be enforced.
B:Where developers get inconsistent coding behavior across branches and members, the engineering workflow can version shared Claude Code guidance alongside the project and review changes; but the production dependency for Workspace separation for sensitive tasks is isolated identities and explicit workspace inventory.
D:This prerequisite supports a method used when code generation sometimes bypasses mandatory lint and secret scans, because the engineering workflow can run deterministic checks at controlled workflow boundaries; however, the named intervention needs isolated identities and explicit workspace inventory before its protection can be enforced.
Question 380
After two network interruptions during an energy-operations pilot, a change request proposes removing the Idempotent action boundary from the tool executor. Which class of side-effect failure becomes more likely?
- An exposure in which undifferentiated feedback cannot guide targeted changes
- An exposure in which response-only logs cannot explain the path to an outcome
- An exposure in which accepting plausible-looking prose can break consumers or misroute cases
- An exposure in which a network timeout can trigger repeated external state changes
Explanation
Correct Answer
D:Removing the safeguard exposes a failure in which a network timeout can trigger repeated external state changes. The architecture normally counters this because idempotency keys and deduplication protect side-effecting writes across retries. A compensating design would require stable request identifiers and a deduplication store and evidence from duplicate external writes per retried request.
Incorrect Answers
A:If reviewers only click thumbs-up or thumbs-down without reason codes, undifferentiated feedback cannot guide targeted changes; but this is not the immediate loss associated with omitting Idempotent action boundary: a network timeout can trigger repeated external state changes.
B:When compliance teams cannot reconstruct which tool outputs informed a recommendation, response-only logs cannot explain the path to an outcome; however, the named safeguard addresses the distinct hazard that a network timeout can trigger repeated external state changes.
C:If downstream software requires machine-readable dispositions with fixed enums, accepting plausible-looking prose can break consumers or misroute cases; whereas removing Idempotent action boundary primarily exposes the risk that a network timeout can trigger repeated external state changes.