Question 220
Following a regression in human-resources services, the team establishes that a single default model policy may violate different governance requirements. Which proposed redesign most directly prevents recurrence?
- Prioritize Per-class model policy
- Prioritize Reasoning-capable model for complex synthesis
- Prioritize Empirical model bake-off
- Prioritize Context-capacity fit before model selection
- Prioritize Small-model first-pass routing
- Prioritize Fallback model under failure
Explanation
Correct Answer
A:The remediation acts at the relevant design boundary: the model configuration can set explicit allowed model configurations by data class and purpose. That counteracts the observed hazard that a single default model policy may violate different governance requirements. Rollout should include approved data classification and deployment policy, with regression evidence from policy-compliant invocation rate by workload class.
Incorrect Answers
B:Where multi-document arbitration requires several interdependent inferences, a stronger model can justify tradeoffs across heterogeneous evidence; but this remedy operates on another boundary, not the case where a single default model policy may violate different governance requirements.
C:When generic leaderboard scores may not predict domain performance, the model configuration can evaluate candidate models on identical task data and operating constraints; whereas the human-resources services defect follows the path in which a single default model policy may violate different governance requirements.
D:Although the alternative reduces the risk that nominal context limits alone do not prove reliable use of every passage because the model configuration can test effective context utilization and recall before choosing a model configuration; yet it would not correct the observed weakness: a single default model policy may violate different governance requirements.
E:This addresses the failure where always using the strongest model increases cost and wait time, because a fast adequate tier handles easy cases and escalates only uncertain ones; but this remedy operates on another boundary, not the case where a single default model policy may violate different governance requirements.
F:The alternative reduces the risk that failover to an untested model can change schema and decision quality because the model configuration can route to an approved fallback model with compatible output contracts; however, the human-resources services defect follows the path in which a single default model policy may violate different governance requirements.
Question 221
The contract administration release team needs a numerical gate for Validated input contract. Rather than equating deployment with success, it wants to catch cases in which unvalidated inputs propagate ambiguity into generated actions. Which indicator should it track?
- Evaluate invalid-input rate and successful recovery by input category
- Evaluate recovered workflows without repeated write operations
- Evaluate percentage of decisions with complete trace lineage
- Evaluate duplicate external writes per retried request
Explanation
Correct Answer
A:An appropriate gate measures invalid-input rate and successful recovery by input category. It tests the mechanism in which schema checks and normalization reject malformed inputs before they enter the model context, rather than a cosmetic demonstration. That distinction matters because unvalidated inputs propagate ambiguity into generated actions.
Incorrect Answers
B:Where multi-step reviews lose progress during retries and restarts, a relevant check would establish whether durable state records completed stages and enables safe continuation; however, that proxy would not demonstrate invalid-input rate and successful recovery by input category for the selected intervention.
C:The alternative indicator checks whether stage-level correlation links input versions, retrieval results, model calls and final decisions when compliance teams cannot reconstruct which tool outputs informed a recommendation; but the acceptance gate must instead inspect invalid-input rate and successful recovery by input category to detect the targeted failure.
D:The proposed measure would suit a setting where a retrying agent sometimes creates the same support task twice, since idempotency keys and deduplication protect side-effecting writes across retries; by contrast, the contract administration release needs evidence from invalid-input rate and successful recovery by input category.
Question 222
The defining problem for a manufacturing quality pilot is that development teams connect tools with overlapping credentials and contracts. Which proposed approach addresses it most precisely?
- Adopt Shared MCP connector governance
- Adopt Hook-based policy checks
- Adopt Repository-level project instructions
- Adopt Least-privilege tool permissions
Explanation
Correct Answer
A:The architectural distinction is that the engineering workflow can register reviewed MCP servers with clear ownership and allowed scopes. The decision should be implemented with central connector catalog and credential lifecycle and evaluated against approved connector use and access-policy compliance; otherwise, ad hoc personal connectors produce inconsistent and insecure access.
Incorrect Answers
B:When code generation sometimes bypasses mandatory lint and secret scans, the engineering workflow can run deterministic checks at controlled workflow boundaries; the stated requirement remains unmet: the engineering workflow can register reviewed MCP servers with clear ownership and allowed scopes.
C:For a workload 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; whereas the manufacturing quality case must guard against the outcome that ad hoc personal connectors produce inconsistent and insecure access.
D:When coding agents request shell access beyond the repository task, the engineering workflow can configure approved actions and deny risky operations by default; this rollout still needs the following guarantee: the engineering workflow can register reviewed MCP servers with clear ownership and allowed scopes.
Question 223
An operations analyst for financial operations must evaluate Stratified production-like test set on real tasks. A test must detect whether a high average score can mask catastrophic minority failures, even when outputs appear fluent. Which measure is appropriate?
- Track schema pass rate and malformed payload count
- Track subgroup coverage and performance against known strata
- Track version-correct answers after policy refresh
- Track invariance violations across controlled transformations
Explanation
Correct Answer
B:Tracking subgroup coverage and performance against known strata provides direct evidence that the system delivers its intended effect: the evaluation process can stratify by task type, risk, language and source complexity. The measure is more informative than prompt counts because the failure being measured is that a high average score can mask catastrophic minority failures.
Incorrect Answers
A:The alternative indicator checks whether the evaluation process can validate outputs mechanically on every release candidate when downstream software expects exact structured fields and allowed enum values; but the acceptance gate must instead inspect subgroup coverage and performance against known strata to detect the targeted failure.
C:When policy retrieval quality looks good using documents from the test period, the evaluation process can test on later source changes to measure freshness and drift response; however, that proxy would not demonstrate subgroup coverage and performance against known strata for the selected intervention.
D:The alternative indicator checks whether the evaluation process can apply controlled input transformations and compare invariant outcomes when answers should not change when irrelevant formatting or names change; but the acceptance gate must instead inspect subgroup coverage and performance against known strata to detect the targeted failure.
Question 224
A logistics coordination incident exposes a specific weakness: prompt rules alone cannot serve as a security boundary. Which architectural correction addresses the cause rather than masking the symptom?
- Enforce Per-user delegated authorization
- Enforce Authenticated webhook processing
- Enforce Short-lived scoped service credentials
- Enforce Server-side tool authorization
- Enforce Tenant-bound retrieval filtering
Explanation
Correct Answer
D:This is the causal correction: the executor evaluates policy before invoking protected downstream APIs. With trusted auth context and policy engine in place, the team can test blocked unauthorized operations in adversarial tool-call tests and show that the problematic case in which prompt rules alone cannot serve as a security boundary has been addressed.
Incorrect Answers
A:Although the alternative reduces the risk that a shared service token can reveal records across users because the integration can enforce each call using the requesting user's authorized scope; but this remedy operates on another boundary, not the case where prompt rules alone cannot serve as a security boundary.
B:Where an upstream application posts actions to a model-driven workflow, the integration can verify message authenticity, timestamp and replay protection before accepting commands; whereas the logistics coordination defect follows the path in which prompt rules alone cannot serve as a security boundary.
C:This addresses the failure where long-lived broad secrets amplify compromise impact, because the integration can issue narrowly scoped renewable credentials for the job identity; yet it would not correct the observed weakness: prompt rules alone cannot serve as a security boundary.
E:This addresses the failure where post-generation redaction cannot undo a cross-tenant retrieval leak, because the integration can apply tenant and document ACL filters before results reach the model; whereas the logistics coordination defect follows the path in which prompt rules alone cannot serve as a security boundary.
Question 225
A human-resources services implementation team has approved Acceptance criteria negotiation. Its expected behavior is that the delivery team can turn goals into testable task-specific, risk-sensitive success criteria. What prerequisite belongs on the release-readiness checklist?
- Prepare measurable quality attributes and named owners
- Prepare accessible data catalog and source stewards
- Prepare scoring rubric and representatives from delivery and risk teams
- Prepare baseline metrics and defined evaluation sample
Explanation
Correct Answer
D:What makes the control operable is baseline metrics and defined evaluation sample. It supplies the precondition for the mechanism whereby the delivery team can turn goals into testable task-specific, risk-sensitive success criteria. The residual failure to guard against is that open-ended satisfaction language cannot determine release readiness, measured through approved acceptance conditions with measurable targets.
Incorrect Answers
A:For the separate challenge where a proposal lists features but omits response time and recovery requirements, the delivery team can capture distinct behavior, latency, security, availability and data constraints; for this release, the team must first verify baseline metrics and defined evaluation sample.
B:For the separate challenge where discovery assumes reliable source records that are fragmented or stale, the delivery team can validate quality, permissions, update cadence and ownership before promising RAG results; however, the named intervention needs baseline metrics and defined evaluation sample before its protection can be enforced.
C:Where teams propose many attractive pilots without common selection criteria, the delivery team can jointly rank candidates by value, feasibility, harm and data readiness; but the production dependency for Acceptance criteria negotiation is baseline metrics and defined evaluation sample.
Question 226
The contract administration go-live board accepts the premise that the control layer can combine prevention, detection, containment and audited recovery. Which preparation is indispensable before it can rely on Defense-in-depth incident response operationally?
- Verify central policy service and auditable subject identity
- Verify layered controls and assigned incident owners
- Verify recipient entitlement map and redaction tests
- Verify parsing boundary and policy enforcement outside the model
- Verify reviewer authority, timeout path and approval evidence
- Verify authenticated tenant attribution and usage telemetry
Explanation
Correct Answer
B:Readiness depends on layered controls and assigned incident owners, not just a design approval. That condition supports the mechanism in which the control layer can combine prevention, detection, containment and audited recovery. Without it, a single classifier cannot handle every emergent attack pattern; monitoring attack containment time and residual failure severity helps verify the deployed result.
Incorrect Answers
A:For the separate challenge where assistants propose privileged write operations in customer systems, the control layer can require server-side authorization and business-policy checks before each side effect; however, the named intervention needs layered controls and assigned incident owners before its protection can be enforced.
C:For the separate challenge where generated summaries may echo sensitive identifiers unnecessary for recipients, the control layer can enforce recipient-specific redaction and validated output fields; for this release, the team must first verify layered controls and assigned incident owners.
D:Where supplier documents include embedded instructions masquerading as system messages, the control layer can treat retrieved text as evidence and prohibit it from changing tool policy; however, the named intervention needs layered controls and assigned incident owners before its protection can be enforced.
E:This prerequisite supports a method used when proposed external transfers could produce irreversible customer harm, because the control layer can stop before execution and require authorized human confirmation; but the production dependency for Defense-in-depth incident response is layered controls and assigned incident owners.
F:For the separate challenge where one customer can monopolize shared model capacity through abusive retries, the control layer can enforce bounded usage budgets and anomaly controls per tenant; for this release, the team must first verify layered controls and assigned incident owners.
Question 227
Following a regression in compliance assurance, the team establishes that unvalidated inputs propagate ambiguity into generated actions. Which proposed redesign most directly prevents recurrence?
- Enforce Idempotent action boundary
- Enforce Audit-correlated processing stages
- Enforce Typed output validation and repair route
- Enforce Validated input contract
Explanation
Correct Answer
D:This intervention addresses the failure at its source because schema checks and normalization reject malformed inputs before they enter the model context. To keep the issue from recurring, establish an explicit schema and ownership of input validation and verify invalid-input rate and successful recovery by input category; otherwise unvalidated inputs propagate ambiguity into generated actions can persist.
Incorrect Answers
A:Although the alternative reduces the risk that a network timeout can trigger repeated external state changes because idempotency keys and deduplication protect side-effecting writes across retries; but this remedy operates on another boundary, not the case where unvalidated inputs propagate ambiguity into generated actions.
B:When response-only logs cannot explain the path to an outcome, stage-level correlation links input versions, retrieval results, model calls and final decisions; whereas the compliance assurance defect follows the path in which unvalidated inputs propagate ambiguity into generated actions.
C:Although 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; yet it would not correct the observed weakness: unvalidated inputs propagate ambiguity into generated actions.
Question 228
A cost-cutting proposal for travel operations eliminates Per-class model policy. Which undesirable consequence would become most credible after that change?
- The possibility that a lightweight classifier can omit crucial cross-document reasoning
- The possibility that always using the strongest model increases cost and wait time
- The possibility that generic leaderboard scores may not predict domain performance
- The possibility that nominal context limits alone do not prove reliable use of every passage
- The possibility that a single default model policy may violate different governance requirements
- The possibility that failover to an untested model can change schema and decision quality
Explanation
Correct Answer
E:The loss of the control permits the situation where a single default model policy may violate different governance requirements. Its governing mechanism is that the model configuration can set explicit allowed model configurations by data class and purpose; without an equivalent mechanism, a favorable demonstration cannot establish safety or quality.
Incorrect Answers
A:When multi-document arbitration requires several interdependent inferences, a lightweight classifier can omit crucial cross-document reasoning; however, the named safeguard addresses the distinct hazard that a single default model policy may violate different governance requirements.
B:If most messages are short and low-complexity while a minority need deep analysis, always using the strongest model increases cost and wait time; whereas removing Per-class model policy primarily exposes the risk that a single default model policy may violate different governance requirements.
C:If stakeholders choose a model based on brand rather than target workload, generic leaderboard scores may not predict domain performance; but this is not the immediate loss associated with omitting Per-class model policy: a single default model policy may violate different governance requirements.
D:The competing hazard is that nominal context limits alone do not prove reliable use of every passage in a workload where legal analysis requests require unusually long source packets; however, the named safeguard addresses the distinct hazard that a single default model policy may violate different governance requirements.
F:The other exposure arises when primary inference service can time out during a critical support window: failover to an untested model can change schema and decision quality; but this is not the immediate loss associated with omitting Per-class model policy: a single default model policy may violate different governance requirements.
Question 229
The post-incident analysis for procurement services concludes that a high average score can mask catastrophic minority failures. Which remediation targets the relevant failure boundary?
- Introduce Stratified production-like test set
- Introduce Metamorphic consistency tests
- Introduce Adversarial injection fixtures
- Introduce Automated contract test suite
Explanation
Correct Answer
A:The reported failure calls for the control under which the evaluation process can stratify by task type, risk, language and source complexity. A prompt-only reminder or unrelated metric would not address the risk that a high average score can mask catastrophic minority failures. Evidence of resolution requires privacy-cleared examples and a sampling design and an improvement in subgroup coverage and performance against known strata.
Incorrect Answers
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; but this remedy operates on another boundary, not the case where a high average score can mask catastrophic minority failures.
C:Where a tool agent ingests external documents that can contain hostile instructions, the evaluation process can add attacker-controlled content and escalation attempts to evaluation; whereas the procurement services defect follows the path in which a high average score can mask catastrophic minority failures.
D:This addresses the failure where manual spot checks miss parser-breaking regressions, because the evaluation process can validate outputs mechanically on every release candidate; yet it would not correct the observed weakness: a high average score can mask catastrophic minority failures.
Question 230
A design authority for field-service coordination has established this requirement: employees connect personal enterprise resources through an integration. Which intervention should be prioritized?
- Introduce OAuth consent and delegated scopes
- Introduce Short-lived scoped service credentials
- Introduce Per-user delegated authorization
- Introduce Tenant-bound retrieval filtering
Explanation
Correct Answer
A:The mechanism is specific: the integration can obtain user consent for minimum scopes and handle revocation. A credible implementation requires supported authorization flow and secure token storage, while its acceptance test monitors revoked-token access denial and scope mismatch rate. The alternative approaches leave the risk that administrator-wide credentials bypass user-controlled consent semantics.
Incorrect Answers
B:When unattended background jobs need limited read access to a narrow dataset, the integration can issue narrowly scoped renewable credentials for the job identity; that does not provide the capability actually needed: the integration can obtain user consent for minimum scopes and handle revocation.
C:When a shared integration must reflect each employee's actual record rights, the integration can enforce each call using the requesting user's authorized scope; the alternative cannot replace this key requirement: the integration can obtain user consent for minimum scopes and handle revocation.
D:When one vector index holds embeddings for several client organizations, the integration can apply tenant and document ACL filters before results reach the model; whereas the field-service coordination case must guard against the outcome that administrator-wide credentials bypass user-controlled consent semantics.