Question 340
A document-extraction service must report the state of every processed page. A conversion worker crashes but a coordinator has cached its last successful output from another document. Which change most directly addresses this condition?
- Return valid page results and flag page nine for targeted correction
- Report passed, failed, and needs-review counts separately
- Return an explicit missing-required-field failure
- Use document identity to prevent reuse of unrelated cached outputs
- Return expected and observed totals with a validation-failure code
Explanation
Correct Answer
D:Cached success for a different input is not a valid substitute; identity-bound artifacts prevent accidental data contamination.
Incorrect Answers
A:For the separate situation where eight pages are parsed, but page nine contains unreadable characters, page-level status avoids losing successful extraction while preventing the unreadable page from being misrepresented as verified; the current task is to reject stale cross-document fallback.
B:Accurate status aggregation ensures that human-review items remain visible instead of being included in successful automation totals; the alternative would make sense if a final report says all invoices passed even though a few were escalated. It does not directly fix the observed condition: a conversion worker crashes but a coordinator has cached its last successful output from another document.
C:This control matters when a required identity field is missing from the extracted record; a ledger-ready record must satisfy required business fields; a syntactically partial result cannot silently pass to commit. The required decision is different: a conversion worker crashes but a coordinator has cached its last successful output from another document.
E:Specific diagnostics help corrective steps distinguish arithmetic conflict from parser failure or missing pages; that assumes a semantic validator detects inconsistent tax arithmetic. The item tests the different ability to reject stale cross-document fallback.
Question 341
A research coordinator must synthesize facts from many long reports. An older report says a rule is active, but a later update supersedes it. What is the most defensible next step?
- Retain a canonical source reference and the needed excerpts once
- Build a short evidence ledger with the relevant claim, number, source, and date
- Store the numeric values and page or figure reference for the chart
- Use effective dates and revision history to resolve precedence
Explanation
Correct Answer
D:Time-sensitive policy questions require temporal reasoning; simple keyword retrieval can privilege an obsolete statement.
Incorrect Answers
A:Repeated text wastes context and can distort apparent evidentiary weight without adding new information; that assumes the prompt is filled with repeated copies of the same source extract. The item tests the different ability to prioritize the current authoritative revision.
B:For the separate situation where a key statistic is buried between hundreds of unrelated facts, a compact ledger brings decisive facts back into active context instead of hoping the writer finds them inside long retrieved text; the current task is to prioritize the current authoritative revision.
C:A lossy narrative can change quantitative meaning; structured values and a locator preserve evidence for verification; the alternative would make sense if a conclusion relies on a small chart extracted from a large PDF. It does not directly fix the observed condition: an older report says a rule is active, but a later update supersedes it.
Question 342
A repository assistant uses several parallel agents to examine unfamiliar services. An analysis worker stalls after requesting another worker recursively. Which approach satisfies the stated constraint?
- Limit task nesting and keep coordination responsibility at the parent
- State the expected test command and success condition in the delegated task
- Return a concise file-indexed summary and paths to relevant excerpts
- Partition write ownership or serialize changes to the shared file
Explanation
Correct Answer
A:Unbounded subagent spawning increases cost and makes failure ownership obscure; a bounded hierarchy is more predictable.
Incorrect Answers
B:A specialist with explicit verification criteria can return evidence of completion instead of an untested code suggestion; that assumes a worker must verify that its proposed fix passes a particular unit test. The item tests the different ability to avoid uncontrolled delegation depth.
C:This control matters when a search agent finishes with fifty raw file excerpts; a narrow handoff saves parent context while retaining references that can be reopened when deeper inspection is needed. The required decision is different: an analysis worker stalls after requesting another worker recursively.
D:For the separate situation where two agents intend to edit the same config file concurrently, parallel reads may be safe, but simultaneous edits to shared state invite races and difficult-to-merge patches; the current task is to avoid uncontrolled delegation depth.
Question 343
A repository team decides how to package MCP integrations for collaboration. A server package is upgraded and its tool arguments change. Which mechanism should be selected?
- Review changed schemas and run integration checks before team rollout
- Use a local-only MCP definition rather than committing it to shared project settings
- Keep the project-scoped server definition under version control without secrets
- Inspect the effective MCP configuration and resolve duplicate definitions
Explanation
Correct Answer
A:Server version changes can break tool calls even when connection setup succeeds, so schema compatibility requires testing.
Incorrect Answers
B:This control matters when a developer tests an experimental local server that colleagues must not use; local scope prevents unreviewed capabilities from silently appearing in every contributor's workspace. The required decision is different: a server package is upgraded and its tool arguments change.
C:For the separate situation where a project-wide tool server has been approved for every contributor to that repository, versioned project configuration gives contributors consistent definitions while leaving credentials outside shared files; the current task is to revalidate clients against the new contract.
D:MCP discovery depends on active connection configuration; changing model instructions cannot fix an incorrect server command; that assumes the same server appears in multiple configuration layers with conflicting commands. The item tests the different ability to revalidate clients against the new contract.
Question 344
A code assistant must choose among repository inspection and execution tools. A test-fix worker has permission to edit only a small package. What should be implemented at the appropriate boundary?
- Give the scanner read access and leave release operations to the pipeline gate
- Invoke the approved test runner and inspect its exit status and output
- Restrict edits and shell working directory to the authorized path
- Use repository search or symbol discovery on the relevant scope
Explanation
Correct Answer
C:Worker-specific file boundaries reduce the risk that a narrowly assigned fix changes unrelated services. For the stated case, the check is whether the system can scope write tools to the owned package; this approach supplies that control.
Incorrect Answers
A:A worker that inspects risk should not control whether risky artifacts are published, especially before independent evaluation; the alternative would make sense if a scanner output must be checked for known-risk entries before publishing. It does not directly fix the observed condition: a test-fix worker has permission to edit only a small package.
B:This control matters when the team asks whether the current test suite passes after a patch; a model-generated prediction cannot verify runtime behavior; the observed test result is the relevant evidence. The required decision is different: a test-fix worker has permission to edit only a small package.
D:For the separate situation where an engineer only asks which files define an interface by name, name-location tasks need index or text inspection; shell execution is unnecessary and may introduce side effects; the current task is to scope write tools to the owned package.
Question 345
A codebase team decides how Claude Code should apply reusable guidance. One engineer wants a shortcut unrelated to the team's shared work. Which option best meets the operational need?
- State the invariant in CLAUDE.md and enforce critical parts in tests
- Store the shortcut in user commands rather than the project
- Create a named project command that accepts the service target
- Implement a hook or CI gate for the required action
- Build a Skill containing an overview and supporting reference files
Explanation
Correct Answer
B:Personal commands avoid cluttering shared team workflows or surprising other contributors. For the stated case, the check is whether the system can use personal-scoped workflow customization; this approach supplies that control.
Incorrect Answers
A:An optional command is not a suitable place for universal code review obligations; the invariant must be visible independently; that assumes an invariant requirement applies to all code edits regardless of command selection. The item tests the different ability to use personal-scoped workflow customization.
C:For the separate situation where developers need to run an audit checklist on demand with a target service, an on-demand workflow supports deliberate audits without making every conversation execute the checklist; the current task is to use personal-scoped workflow customization.
D:Prompt or command availability does not guarantee execution; event-based or pipeline enforcement closes that gap; the alternative would make sense if a workflow must run even when Claude chooses not to call a named command. It does not directly fix the observed condition: one engineer wants a shortcut unrelated to the team's shared work.
E:This control matters when a specialist procedure involves a folder of reference templates and examples; a Skill can organize larger task-specific knowledge while keeping routine sessions from carrying all of it. The required decision is different: one engineer wants a shortcut unrelated to the team's shared work.
Question 346
A customer-service agent uses hooks around tool operations. A prompt says 'never run shell commands,' but the tool remains available and callable. What is the most appropriate design choice?
- Transform dates at a controlled post-tool boundary into a canonical format
- Record the event at the tool-execution boundary using a hook or service log
- Remove or deny shell execution in the actual tool permissions or hook
- Use a pre-execution hook or service policy that rejects unauthorized refund calls
Explanation
Correct Answer
C:Text guidance is probabilistic, whereas execution controls can reject a forbidden operation regardless of model output.
Incorrect Answers
A:This control matters when a source-search tool returns inconsistent date formats that confuse downstream parsing; a post-tool normalization step makes structured evidence consistent without hoping a later prompt corrects every variant. The required decision is different: a prompt says 'never run shell commands,' but the tool remains available and callable.
B:Execution-time logging captures real calls and outcomes; model-generated explanations alone are not reliable audit evidence; that assumes the application needs a durable record whenever a privileged tool is invoked. The item tests the different ability to make the tool restriction enforceable.
D:For the separate situation where a refund tool must never execute for amounts above the permitted threshold without approval, a deterministic tool boundary can block a high-value mutation even if a model instruction fails to deter it; the current task is to make the tool restriction enforceable.
Question 347
An application extracts structured fields from invoices with Claude. The system stores several line items for one invoice. Which response is most reliable?
- Declare amount as a number and reject nonnumeric values
- Define an array of line-item objects with required amount and description fields
- Have Claude supply fields through a tool interface with a defined JSON Schema
- Model tax ID as optional or nullable according to downstream semantics
Explanation
Correct Answer
B:Nested arrays preserve line-item boundaries and permit validation of each element without relying on free-text delimiters.
Incorrect Answers
A:Type constraints prevent textual approximations from being silently interpreted as ledger-ready amounts; the alternative would make sense if a paid amount is emitted as a phrase such as 'about fifty'. It does not directly fix the observed condition: the system stores several line items for one invoice.
C:For the separate situation where free-form JSON text frequently contains trailing commentary that breaks parsing, tool-argument schemas provide a clearer machine-readable contract than merely asking for 'JSON only' in prose, though validation remains necessary; the current task is to use a typed repeated-item structure.
D:A schema that requires nonexistent information encourages fabricated values; absence must be representable explicitly; that assumes some invoice records omit an optional tax identifier. The item tests the different ability to use a typed repeated-item structure.
Question 348
A support assistant assigns risk and confidence to proposed customer actions. A high-impact refund crosses a configured approval threshold. What should the team do to resolve this particular gap?
- Compare confidence buckets against observed correctness and adjust routing thresholds
- Require actual policy checks and relevant source fields before approving action
- Route the transaction through the mandatory human approval gate
- Request the missing field or hand off instead of guessing
Explanation
Correct Answer
C:The threshold is an execution policy independent of the model's expressed confidence about the refund. For the stated case, the check is whether the system can require review regardless of model assurance; this approach supplies that control.
Incorrect Answers
A:Reliability requires empirical error measurement; a fluent explanation is not proof that reported certainty matches accuracy; the alternative would make sense if monitoring shows confident unsupported answers increasing after a prompt change. It does not directly fix the observed condition: a high-impact refund crosses a configured approval threshold.
B:For the separate situation where a high-confidence model verdict has no linked eligibility evidence, confidence scores are not evidence; external authorization and validated facts determine whether a sensitive operation may proceed; the current task is to require review regardless of model assurance.
D:This control matters when a decision has low confidence because required account information is missing; uncertainty arising from missing information needs a data remedy, not stronger prompt language or fabricated certainty. The required decision is different: a high-impact refund crosses a configured approval threshold.
Question 349
A document-extraction pipeline exposes conversion and validation tools to Claude. An address verifier uses an external data source that may be outdated. Which control is most directly relevant?
- Constrain the input schema to supported currency codes and document normalization
- Split distinct operations into purpose-built tools with focused schemas
- Return page identifiers, text, status, and errors as separate structured records
- Include lookup time and source provenance in the result contract
Explanation
Correct Answer
D:A timestamped external check lets downstream logic decide whether the verification is fresh enough for the task. Here it serves the explicit need to disclose temporal limits of verification without adding a separate unverified inference.
Incorrect Answers
A:For the separate situation where a validator requires normalized currency codes but callers send symbols, a typed canonical contract prevents ambiguous symbols from becoming silently incorrect financial records; the current task is to disclose temporal limits of verification.
B:Narrow functions are easier for Claude to select and for the host to validate than a parameter-heavy command that mixes unrelated side effects; the alternative would make sense if the team continually adds loosely related parameters to one giant function. It does not directly fix the observed condition: an address verifier uses an external data source that may be outdated.
C:This control matters when an OCR tool can return multiple pages with partial failures; per-page status distinguishes a partly readable document from a fully successful extraction and supports targeted retries. The required decision is different: an address verifier uses an external data source that may be outdated.
Question 350
An invoice-extraction workflow must recover from malformed output. The model has failed the same correction twice. How should the workflow be adjusted?
- Stop after the configured retry limit and return a review-needed state
- Report the mismatched field and constraint, then retry within a bounded policy
- Require evidence or field-level source matching for supplier identity
- Run business-rule arithmetic validation and request targeted correction
- Mark the conflict and request human clarification with both source locations
Explanation
Correct Answer
A:Unlimited repair attempts can waste resources and may still fabricate a plausible record; bounded retries preserve reliability.
Incorrect Answers
B:For the separate situation where a tool payload is not valid against the required JSON Schema, specific schema errors give the model a repair target; blindly regenerating without feedback repeats the same failure mode; the current task is to bound the repair loop and surface failure.
C:Syntax repair does not imply factual correctness; source-backed validation checks whether each critical value is actually present; the alternative would make sense if a corrected output has valid JSON but an invented supplier name. It does not directly fix the observed condition: the model has failed the same correction twice.
D:This control matters when an extracted total passes type validation but does not equal subtotal plus taxes; a valid numeric type does not prove accounting consistency; arithmetic rules catch errors outside the schema's expressiveness. The required decision is different: the model has failed the same correction twice.
E:A retry cannot resolve contradictory evidence by itself; the ambiguity belongs in review rather than arbitrary selection; that assumes the invoice number appears in two places with different values. The item tests the different ability to bound the repair loop and surface failure.