Question 140
A compliance manual in Project knowledge was replaced centrally, but the Project still holds an older uploaded copy and gives obsolete steps. What is the most defensible approach?
- Refresh an uploaded file after a controlled update
- Check underlying permission when a connector returns no document
- Audit connector usefulness before keeping it enabled
- Avoid duplicating documents with conflicting titles
- Give broad access to all organizational connectors by default
- Place the recurring rule only in one transient conversation for the shared Project
Explanation
Correct Answer
A:Uploaded material needs lifecycle maintenance. Replace the old copy and re-test prompts referencing changed sections.
Incorrect Answers
B:Connectors operate within authorized access boundaries; the missing result may reflect permissions rather than a model failure; this helps with distinguishing retrieval authorization from generation quality while leaving keeping static knowledge synchronized with revisions unsolved.
C:Remove unnecessary integrations and periodically check owner and purpose, while the requirement here is keeping static knowledge synchronized with revisions instead of minimizing unused connector exposure.
D:Duplicate overlapping references can create ambiguity; maintain a single controlled document or explicit precedence rules; that distinction matters because the specific gap is keeping static knowledge synchronized with revisions, not resolving conflicts among repeated knowledge files.
E:Unnecessary connections expand accessible data and ignore least-privilege and data-policy constraints, which does not replace the required process of keeping static knowledge synchronized with revisions.
F:Chat-level guidance does not automatically become persistent instructions for every new conversation in a Project; the actual task calls for keeping static knowledge synchronized with revisions instead.
Question 141
A project team wants to connect a customer-support system to Claude without identifying which records contain payment details or protected conversations. What is the most appropriate next action?
- Aggregate customer-support tickets before assessing whether protected details may be accessed
- Strip unique identifiers not needed for the question
- Classify data before connecting a new source
- Honor a contractual restriction on third-party processing
Explanation
Correct Answer
C:Determine sensitivity and permitted processing before enabling access. Connector availability is not blanket data authorization.
Incorrect Answers
A:Aggregation can reduce personal-data exposure when designed to prevent disclosure of small groups, while the requirement here is assessing source sensitivity before access instead of reducing person-level detail in analytical inputs.
B:Only collect or process fields necessary for the approved purpose; unused sensitive columns add risk without analytical benefit; this helps with minimizing unnecessary identity attributes while leaving assessing source sensitivity before access unsolved.
D:Contractual controls define what processing is permitted; obtain appropriate approval or use a compliant alternative; prompt wording cannot waive the restriction; that distinction matters because the specific gap is assessing source sensitivity before access, not respecting external processing restrictions.
Question 142
An instruction change fixes the missing risk paragraph but may have removed mandatory owner fields from a recurring report. Which action would address the stated problem most directly?
- Re-run a small acceptance suite after prompt edits
- Write a change-note explaining the fixed risk paragraph
- Update the core template for a recurring issue
- Preserve accepted factual content during a tone revision
Explanation
Correct Answer
A:Regression cases check both targeted fixes and unaffected requirements. Prompt success must be measured across representative tasks.
Incorrect Answers
B:Link each change to a reproducible issue and verification result; this supports evidence-based iteration and rollback; this helps with tracing improvements to observed failures while leaving verifying that repairs did not break other behavior unsolved.
C:Fix the persistent specification rather than depend on repeated follow-ups, yet the scenario asks for verifying that repairs did not break other behavior rather than eliminating a repeated instruction omission.
D:Scope editing to voice and wording while explicitly retaining approved dates, amounts, and obligations, while the requirement here is verifying that repairs did not break other behavior instead of protecting checked facts during editorial change.
Question 143
An incident report mixes timestamps with participants' opinions, and the review team needs an unbiased timeline before it discusses causes. What should the team do next?
- Circulate the current incident draft for sign-off before separating chronology from opinion
- Separate extraction of facts from interpretation
- Design a data-cleaning step before calculating metrics
- Build a preliminary rubric before evaluating submissions
Explanation
Correct Answer
B:Extracting verifiable events first creates an evidentiary base. Interpretation can then be labeled and checked separately instead of contaminating the timeline.
Incorrect Answers
A:A pilot review with users reveals operational gaps before the script becomes standard; this helps with testing a proposed process with actual users while leaving distinguishing events from causal judgments unsolved.
C:Cleaning and defining records first stops duplicate observations and malformed dates from corrupting the metric; that is relevant to fixing record quality before calculation, not the stated need for distinguishing events from causal judgments.
D:An agreed rubric aligns fields, scoring definitions, and exclusions across all submissions, a useful capability for standardizing evaluation dimensions ahead of review but not a solution to distinguishing events from causal judgments.
Question 144
A generated event plan assumes every attendee can use stairs despite the facilities brief explicitly including accessibility constraints. Which option best resolves the stated issue?
- Inspect whether the model selected evidence favoring the easiest access route
- Check omission of accessibility needs
- Validate that an attribution is not composite
- Investigate fabricated meeting commitments
Explanation
Correct Answer
B:A useful plan must account for affected user groups rather than silently adopting a narrow default. Compare the proposal with stated access requirements.
Incorrect Answers
A:Leading instructions distort evidence selection; ask for a neutral assessment with supportive and adverse indicators and consistent thresholds; that distinction matters because the specific gap is including accessibility requirements, not preventing a leading prompt from skewing evaluation.
C:A quotation must trace to one actual speaker and passage; composite paraphrases should not be presented as verbatim attributable speech, while the requirement here is including accessibility requirements instead of preventing misattributed composite quotes.
D:Decision extraction must distinguish proposals, discussion, approvals, and follow-ups; this helps with distinguishing proposals from recorded decisions while leaving including accessibility requirements unsolved.
Question 145
A coordinator sends hundreds of short independent classifications separately and the workflow can safely process them in organized groups. Which decision is the strongest fit for this task?
- Send every routine independent classification to an expert for manual handling
- Use a small validated model for simple extraction
- Batch similar tasks where practical
- Measure error-adjusted cost rather than price per request only
Explanation
Correct Answer
C:Grouping similar work reduces process overhead while keeping task boundaries and traceability. Batch size should preserve quality and auditability.
Incorrect Answers
A:Quality controls should be risk-weighted rather than eliminated for cost savings; this helps with protecting high-impact review gates while leaving improving throughput without losing record boundaries unsolved.
B:Routine extraction may not require the highest reasoning tier; validate against representative forms and track exceptions, yet the scenario asks for improving throughput without losing record boundaries rather than reducing model expense on well-defined extraction.
D:Total workflow cost includes human rework and rejected outputs; per-request price alone can misrepresent economic value, while the requirement here is improving throughput without losing record boundaries instead of comparing end-to-end work cost.
Question 146
A requirements document says the assistant should 'respond helpfully' but offers no test for cited accuracy or proper escalation. Which option best resolves the stated issue?
- Prototype a screen concept as an Artifact
- Release consequential decisions based on the draft alone during workflow design
- Escalate production API and identity architecture
- Draft acceptance tests for a planned solution
- Design a human approval checkpoint
Explanation
Correct Answer
D:Acceptance tests turn ambitions into observable input-output cases. They guide implementation review without assuming success.
Incorrect Answers
A:A lightweight prototype makes layout and decision flow tangible for stakeholder feedback, while the requirement here is defining testable behavior for a solution instead of exploring interface usability before development.
B:Unreviewed consequential decisions remove accountability and do not address error and bias risks, which does not replace the required process of defining testable behavior for a solution.
C:Authentication, authorization, and failure recovery need engineering design and organizational controls; this helps with routing complex systems engineering to specialists while leaving defining testable behavior for a solution unsolved.
E:An authorized reviewer at the publication boundary controls external commitments, yet the scenario asks for defining testable behavior for a solution rather than keeping approval before external release.
Question 147
Project instructions say to follow a current contract, while an older FAQ contains a conflicting rate and Claude alternates between the two. What is the most appropriate next action?
- Rewrite the house-style guidance
- Define source precedence explicitly
- State escalation conditions for out-of-scope requests
- Describe role, task and output expectations
Explanation
Correct Answer
B:Priority rules should name the governing source and behavior on conflict. Remove obsolete information when possible rather than depend on silent guessing.
Incorrect Answers
A:Coherent instructions avoid contradictory tone rules; resolve conflicts based on the actual audience and purpose; that is relevant to removing inconsistent style directives, not the stated need for resolving which instruction or source controls.
C:Instructions can define approval thresholds and when to refer a case to a manager, while the requirement here is resolving which instruction or source controls instead of identifying the point where a human must authorize.
D:Specific role, audience, format, and sourcing rules make behavior testable, a useful capability for creating actionable reusable behavioral guidance but not a solution to resolving which instruction or source controls.
Question 148
An employee wants to keep generated summaries of sensitive investigations indefinitely without checking organizational record-keeping requirements. Which method best fits these constraints?
- Add a conversational promise of confidentiality
- Confirm permitted cross-border processing
- Check indirect reidentification in supposedly anonymous data
- Apply the relevant retention and records policy
Explanation
Correct Answer
D:Retention, deletion, and access obligations may apply to both input and output. Follow approved records controls.
Incorrect Answers
A:A request in the prompt does not establish contractual, account, or system-level handling guarantees, while the requirement here is managing sensitive generated records across their lifecycle instead of using enforceable privacy mechanisms.
B:Authorization must include applicable transfer and processing constraints; that is relevant to validating geographic processing obligations, not the stated need for managing sensitive generated records across their lifecycle.
C:Combinations of quasi-identifiers can reveal a person; evaluate linkage risk before publication or AI processing, a useful capability for preventing identity inference from sparse records but not a solution to managing sensitive generated records across their lifecycle.
Question 149
An editor needs to choose between concise and explanatory customer messages but repeated 'try again' requests yield nearly identical drafts. Which decision is the strongest fit for this task?
- Stop using the editing workflow after a single unsatisfactory draft version
- Ask for alternate versions with explicit differences
- Escalate when a prompt fix cannot grant data access
- Add tests for an omitted exception rule
Explanation
Correct Answer
B:Name the dimensions to vary and compare versions against the same source facts. Variation should serve a review decision.
Incorrect Answers
A:Measured error costs justify redesign or escalation of hard cases; price per call is not the only optimization criterion, a useful capability for correcting a cost-saving method that harms quality but not a solution to producing meaningfully different drafting alternatives.
C:Authorization problems require an approved access change or different workflow; this helps with resolving a system access boundary while leaving producing meaningfully different drafting alternatives unsolved.
D:Encode the exception and test both promoted and ordinary purchases; narrow tests expose whether the rule works across conditions; that is relevant to correcting a missing special-case rule, not the stated need for producing meaningfully different drafting alternatives.
Question 150
After four exchanges, Claude still mixes gross and net revenue because the initial task never defined which metric is required. Which action would address the stated problem most directly?
- Score each revision on a checklist that still leaves gross and net revenue undefined
- Point to the missing evidence and request a correction
- Revise the prompt when its original goal was ambiguous
- Request a substantive self-critique against criteria
Explanation
Correct Answer
C:Repairing the original definition removes a systematic ambiguity from the task. Endless follow-ups leave conflicting instructions in context.
Incorrect Answers
A:A checklist catches mandatory omissions independently of how polished the text becomes, a useful capability for verifying that mandatory content survived revision but not a solution to correcting an ambiguous initial metric.
B:Targeted feedback directs repair to identifiable claims and asks for verifiable support, yet the scenario asks for correcting an ambiguous initial metric rather than repairing specifically unsupported claims.
D:A structured critique can surface weaknesses for follow-up; it is a review aid rather than independent proof that the draft is accurate; that distinction matters because the specific gap is correcting an ambiguous initial metric, not identifying risks for subsequent human checking.