Question 510
A team maintaining a secret rotation incident must meet 3 separate acceptance criteria: (1) stop frontend users from receiving unrestricted backend tool credentials; (2) stop an authenticated user from reading another tenant's records via a tool; (3) stop a coding agent from running commands that match a prohibited destructive pattern. Which three implementation actions directly satisfy the stated requirements? Choose THREE.
- Check resource-level authorization for every sensitive tool request.
- Verify origin and integrity of sensitive external data where supported.
- Enforce a pre-tool hook that denies the disallowed call.
- Use a server-side workload identity or scoped service credential.
- Isolate external search results and validate any resulting tool action independently.
- Remove or mask unnecessary personal information before model submission.
Explanation
Correct Answers
A:The test to stop an authenticated user from reading another tenant's records via a tool is met because proving identity does not grant access to every object; the backend must enforce tenant and resource permissions separately; the remaining selections covers the separate need to stop a coding agent from running commands that match a prohibited destructive pattern and need to stop frontend users from receiving unrestricted backend tool credentials.
C:The test to stop a coding agent from running commands that match a prohibited destructive pattern is met because a deterministic hook checks the proposed command before execution, providing a stronger control than asking the agent to self-police; the remaining selections covers the separate need to stop an authenticated user from reading another tenant's records via a tool and need to stop frontend users from receiving unrestricted backend tool credentials.
D:The test to stop frontend users from receiving unrestricted backend tool credentials is met because service identities tie privileges to a runtime workload and avoid exposing long-lived secrets to browsers or untrusted users; the remaining selections covers the separate need to stop an authenticated user from reading another tenant's records via a tool and need to stop a coding agent from running commands that match a prohibited destructive pattern.
Incorrect Answers
B:The action is useful for the separate capability to detect unauthorized modifications to a tool result before it reaches the model because trusted decisions require confidence that tool outputs were not altered; transport and signature checks complement content validation; however, the required test outcomes are to stop an authenticated user from reading another tenant's records via a tool and to stop a coding agent from running commands that match a prohibited destructive pattern and to stop frontend users from receiving unrestricted backend tool credentials.
E:The action is useful for the separate capability to reject a malicious instruction inserted into a web-search result rather than the user's direct request because indirect attacks exploit lower-trust retrieval surfaces, so trusting search snippets as operational directives crosses a security boundary; however, the required test outcomes are to stop an authenticated user from reading another tenant's records via a tool and to stop a coding agent from running commands that match a prohibited destructive pattern and to stop frontend users from receiving unrestricted backend tool credentials.
F:The action is useful for the separate capability to protect customer identity fields when only document topics are needed for classification because minimization limits exposure while preserving task inputs; relying on the model to ignore visible PII does not provide the same protection; however, the required test outcomes are to stop an authenticated user from reading another tenant's records via a tool and to stop a coding agent from running commands that match a prohibited destructive pattern and to stop frontend users from receiving unrestricted backend tool credentials.
Question 511
A production readiness review of a structured output service finds that the team cannot yet ask for missing essential parameters rather than inventing a consequential choice or maintain label semantics when a few-shot prompt contains dialogue. Which paired changes directly remedy the two gaps?
- Perform both steps: (1) Instruct the application to request clarification for unresolved high-impact ambiguity; (2) use properly labeled example inputs and outputs instead of mixing speakers.
- Perform both steps: (1) Ground consequential statements in retrieved evidence or external checks; (2) use properly labeled example inputs and outputs instead of mixing speakers.
- Perform both steps: (1) Ground consequential statements in retrieved evidence or external checks; (2) name each mandatory field or section and provide formatting constraints.
- Perform both steps: (1) Instruct the application to request clarification for unresolved high-impact ambiguity; (2) name each mandatory field or section and provide formatting constraints.
Explanation
Correct Answer
A:The first change enables the team to ask for missing essential parameters rather than inventing a consequential choice because clarification reduces irreversible guesses when requirements do not uniquely determine a safe action; the second addresses the separate need to maintain label semantics when a few-shot prompt contains dialogue because correct role boundaries keep demonstrations interpretable and prevent an example answer from looking like a live instruction.
Incorrect Answers
B:The second step meets the requirement to maintain label semantics when a few-shot prompt contains dialogue because correct role boundaries keep demonstrations interpretable and prevent an example answer from looking like a live instruction; the first step would only help to avoid treating fluent model wording as evidence that a claim is verified because linguistic confidence is not a calibrated truth guarantee, so factual correctness requires independent support, leaving the separate requirement to ask for missing essential parameters rather than inventing a consequential choice unmet.
C:Both steps target different results—avoid treating fluent model wording as evidence that a claim is verified and require that a generated report have a fixed set of sections—because linguistic confidence is not a calibrated truth guarantee, so factual correctness requires independent support and concrete desired structure is easier to verify than a vague instruction to be professional or comprehensive; the required capabilities are instead to ask for missing essential parameters rather than inventing a consequential choice and to maintain label semantics when a few-shot prompt contains dialogue.
D:The first step is appropriate for the requirement to ask for missing essential parameters rather than inventing a consequential choice because clarification reduces irreversible guesses when requirements do not uniquely determine a safe action; the other step instead aims to require that a generated report have a fixed set of sections because concrete desired structure is easier to verify than a vague instruction to be professional or comprehensive, which does not meet the second requirement to maintain label semantics when a few-shot prompt contains dialogue.
Question 512
For a multi-stage support automation, a design is acceptable only if it can avoid adding orchestration infrastructure to a single request-and-response feature while also being able to stop two parallel branches from overwriting the same mutable record. Which paired implementation is sufficient?
- Perform both steps: (1) Prefer a simple API call over a general-purpose multi-agent framework; (2) use isolated branch state and a controlled reconciliation step.
- Perform both steps: (1) Prefer a simple API call over a general-purpose multi-agent framework; (2) consume the agent runtime's streamed event sequence.
- Perform both steps: (1) Configure tool permissions rather than depending only on a prompt; (2) consume the agent runtime's streamed event sequence.
- Perform both steps: (1) Configure tool permissions rather than depending only on a prompt; (2) use isolated branch state and a controlled reconciliation step.
Explanation
Correct Answer
A:The first change enables the team to avoid adding orchestration infrastructure to a single request-and-response feature because frameworks are useful for genuinely branching, long-lived execution, but increase complexity when no orchestration problem exists; the second addresses the separate need to stop two parallel branches from overwriting the same mutable record because explicit merge rules are required when outputs converge; concurrent mutation without arbitration causes nondeterministic results.
Incorrect Answers
B:The first step is appropriate for the requirement to avoid adding orchestration infrastructure to a single request-and-response feature because frameworks are useful for genuinely branching, long-lived execution, but increase complexity when no orchestration problem exists; the other step instead aims to deliver intermediate agent events to the client as they occur because event-driven consumption lets the application display progress and inspect tool actions without waiting for the entire multi-step run to finish, which does not meet the second requirement to stop two parallel branches from overwriting the same mutable record.
C:Both steps target different results—restrict a development agent to a defined set of file and command capabilities and deliver intermediate agent events to the client as they occur—because runtime permissions constrain what actions can run even if the model requests more access; prompt instructions alone are not a security boundary and event-driven consumption lets the application display progress and inspect tool actions without waiting for the entire multi-step run to finish; the required capabilities are instead to avoid adding orchestration infrastructure to a single request-and-response feature and to stop two parallel branches from overwriting the same mutable record.
D:The second step meets the requirement to stop two parallel branches from overwriting the same mutable record because explicit merge rules are required when outputs converge; concurrent mutation without arbitration causes nondeterministic results; the first step would only help to restrict a development agent to a defined set of file and command capabilities because runtime permissions constrain what actions can run even if the model requests more access; prompt instructions alone are not a security boundary, leaving the separate requirement to avoid adding orchestration infrastructure to a single request-and-response feature unmet.
Question 513
To ship a Claude API workload benchmark, the application must invite stylistic variety during noncritical brainstorming and allocate additional reasoning effort to a complex multistep decision where supported. Which pair of listed actions is necessary for those stated capabilities? Choose TWO.
- Use the model's documented thinking capability within the latency budget.
- Budget total context for instructions, history, tools, and output.
- Use more exploratory sampling where supported and measure quality.
- Benchmark an Opus-family model on the hardest cases.
- Provide one accurate input-output demonstration.
Explanation
Correct Answers
A:To allocate additional reasoning effort to a complex multistep decision where supported as part of a design that must also invite stylistic variety during noncritical brainstorming, reasoning settings can change token consumption and response latency and should be evaluated for the particular workload; this is why the action is one of the two required selections.
C:To invite stylistic variety during noncritical brainstorming as part of a design that must also allocate additional reasoning effort to a complex multistep decision where supported, higher sampling randomness broadens candidate phrasing but can reduce predictability, so it suits ideation better than strict extraction; this is why the action is one of the two required selections.
Incorrect Answers
B:This alternative serves the separate goal to know why a long prompt plus a proposed response may exceed the model's accepted context because a context window bounds what can be processed in one request; large tool results consume it along with user messages; the required outcomes are instead to invite stylistic variety during noncritical brainstorming and to allocate additional reasoning effort to a complex multistep decision where supported.
D:This alternative serves the separate goal to solve demanding multistep reasoning tasks where quality matters more than lowest unit price because higher-capability tiers may improve difficult-task performance, but their cost and latency still need objective justification; the required outcomes are instead to invite stylistic variety during noncritical brainstorming and to allocate additional reasoning effort to a complex multistep decision where supported.
E:This alternative serves the separate goal to teach a distinctive output format with one representative example because a single example illustrates structure and conventions, but it may not reveal how to handle difficult edge cases; the required outcomes are instead to invite stylistic variety during noncritical brainstorming and to allocate additional reasoning effort to a complex multistep decision where supported.
Question 514
In a secure tool-execution boundary, testing reveals that the system cannot demonstrate who approved a privileged tool action and when and cannot capture outcomes of agent command execution for security review. Select the two targeted actions that close these gaps. Choose TWO.
- Record result metadata in a post-tool hook.
- Record authenticated approval events with actor, scope, and request correlation.
- Isolate external search results and validate any resulting tool action independently.
- Combine prompt boundaries, scoped credentials, validation, and approval gates.
- Retrieve keys through a managed secrets facility at runtime.
Explanation
Correct Answers
A:To capture outcomes of agent command execution for security review as part of a design that must also demonstrate who approved a privileged tool action and when, post-execution hooks can log and inspect what happened; they are unsuitable as the sole barrier to a side effect that already occurred; this is why the action is one of the two required selections.
B:To demonstrate who approved a privileged tool action and when as part of a design that must also capture outcomes of agent command execution for security review, an audit trail binds authorization to a specific action, making privilege use reviewable without treating model-generated text as proof of consent; this is why the action is one of the two required selections.
Incorrect Answers
C:This alternative serves the separate goal to reject a malicious instruction inserted into a web-search result rather than the user's direct request because indirect attacks exploit lower-trust retrieval surfaces, so trusting search snippets as operational directives crosses a security boundary; the required outcomes are instead to demonstrate who approved a privileged tool action and when and to capture outcomes of agent command execution for security review.
D:This alternative serves the separate goal to secure an agent against both malicious retrieved text and unsafe tool execution because defense in depth assumes individual filters can fail; independent enforcement boundaries reduce the chance of harmful side effects; the required outcomes are instead to demonstrate who approved a privileged tool action and when and to capture outcomes of agent command execution for security review.
E:This alternative serves the separate goal to keep Claude API credentials out of repository source and deployment images because central secret handling improves rotation and auditing and avoids accidental credential disclosure through source control; the required outcomes are instead to demonstrate who approved a privileged tool action and when and to capture outcomes of agent command execution for security review.
Question 515
While planning a tool-using incident triage agent, the product owner specifies that it must ensure each later tool receives a validated artifact from its preceding stage. Which solution fulfills the requirement?
- Evaluate a managed agent runtime against the deployment requirements.
- Give an agent a bounded objective and a tool-selection loop.
- Fan out to role-specific reviewers and aggregate disagreements.
- Chain dependent operations serially with explicit input and output contracts.
Explanation
Correct Answer
D:To ensure each later tool receives a validated artifact from its preceding stage in this tool-using incident triage agent, later steps cannot safely start until upstream data exists; passing validated outputs prevents races and malformed downstream input; the proposed implementation directly supplies that capability.
Incorrect Answers
A:The proposed alternative helps to minimize infrastructure ownership for a supported hosted agent execution surface because a managed runtime can take on execution infrastructure, but access controls, product availability, and data-handling needs still require evaluation; that mechanism does not meet the different requirement to ensure each later tool receives a validated artifact from its preceding stage.
B:The proposed alternative helps to pursue an open-ended investigation where the next action depends on intermediate evidence because agentic control lets the model choose among permitted steps as observations arrive; budgets and termination conditions still bound its autonomy; that mechanism does not meet the different requirement to ensure each later tool receives a validated artifact from its preceding stage.
C:The proposed alternative helps to have independent specialized reviewers inspect the same proposal from different perspectives because separating reviews allows conflicting evidence to surface before the supervisor combines findings; a single sequential review may miss independent critiques; that mechanism does not meet the different requirement to ensure each later tool receives a validated artifact from its preceding stage.
Question 516
To ship a TypeScript Claude integration, the application must restore a known-working prompt and model configuration after quality degrades and ensure each production feature can be tied to an approved acceptance test. Which pair of listed actions is necessary for those stated capabilities? Choose TWO.
- Keep versioned artifacts and a tested rollback procedure.
- Maintain links between requirements, implementation changes, and verification cases.
- Treat model migration as a tested deployment with staged traffic.
- Store team settings in the project's shared configuration file.
- Add application-owned approval logic around the refund tool.
Explanation
Correct Answers
A:To restore a known-working prompt and model configuration after quality degrades as part of a design that must also ensure each production feature can be tied to an approved acceptance test, model IDs, prompts, schemas, and tool contracts are release components; restoring just the application binary may not restore prior behavior; this is why the action is one of the two required selections.
B:To ensure each production feature can be tied to an approved acceptance test as part of a design that must also restore a known-working prompt and model configuration after quality degrades, requirement traceability helps prevent a prompt or model update from silently changing behavior outside the agreed product scope; this is why the action is one of the two required selections.
Incorrect Answers
C:This alternative serves the separate goal to compare candidate models before switching a mission-critical endpoint because model upgrades may alter output distributions or supported features; staged validation reduces the chance of an unnoticed behavior shift; the required outcomes are instead to restore a known-working prompt and model configuration after quality degrades and to ensure each production feature can be tied to an approved acceptance test.
D:This alternative serves the separate goal to apply consistent reviewed tool permissions to an entire Claude Code repository team because repository-level settings can be versioned and reviewed, unlike private preferences that vary across each developer's machine; the required outcomes are instead to restore a known-working prompt and model configuration after quality degrades and to ensure each production feature can be tied to an approved acceptance test.
E:This alternative serves the separate goal to ensure refunds above a policy limit receive human sign-off because decisions that create financial side effects require deterministic authorization checks beyond a model's recommendation; the required outcomes are instead to restore a known-working prompt and model configuration after quality degrades and to ensure each production feature can be tied to an approved acceptance test.
Question 517
Two acceptance tests block a latency-sensitive model gateway: one requires the ability to estimate model usage from encoded text rather than counting words; the other requires the ability to teach a distinctive output format with one representative example. Which plan clears both tests?
- Perform both steps: (1) Measure or count input and output tokens with the supported tokenizer or API; (2) measure cache-write overhead against subsequent cache-read savings.
- Perform both steps: (1) Model the response as conditioned token generation; (2) provide one accurate input-output demonstration.
- Perform both steps: (1) Measure or count input and output tokens with the supported tokenizer or API; (2) provide one accurate input-output demonstration.
- Perform both steps: (1) Model the response as conditioned token generation; (2) measure cache-write overhead against subsequent cache-read savings.
Explanation
Correct Answer
C:The first change enables the team to estimate model usage from encoded text rather than counting words because billing and context consumption track tokens, which do not map one-to-one to English words or characters; the second addresses the separate need to teach a distinctive output format with one representative example because a single example illustrates structure and conventions, but it may not reveal how to handle difficult edge cases.
Incorrect Answers
A:The first step is appropriate for the requirement to estimate model usage from encoded text rather than counting words because billing and context consumption track tokens, which do not map one-to-one to English words or characters; the other step instead aims to reduce repeated cost for a stable lengthy policy prefix reused many times because cache writes may cost extra while hits are cheaper; the reuse frequency determines whether the optimization saves money, which does not meet the second requirement to teach a distinctive output format with one representative example.
B:The second step meets the requirement to teach a distinctive output format with one representative example because a single example illustrates structure and conventions, but it may not reveal how to handle difficult edge cases; the first step would only help to explain why language-model output is generated step by step rather than fetched from a fixed response table because autoregressive decoding selects subsequent tokens based on supplied context and prior output, which creates flexible but non-guaranteed wording, leaving the separate requirement to estimate model usage from encoded text rather than counting words unmet.
D:Both steps target different results—explain why language-model output is generated step by step rather than fetched from a fixed response table and reduce repeated cost for a stable lengthy policy prefix reused many times—because autoregressive decoding selects subsequent tokens based on supplied context and prior output, which creates flexible but non-guaranteed wording and cache writes may cost extra while hits are cheaper; the reuse frequency determines whether the optimization saves money; the required capabilities are instead to estimate model usage from encoded text rather than counting words and to teach a distinctive output format with one representative example.
Question 518
An end-to-end test for an operations research assistant checks whether it can coordinate several specialized investigations and assemble one accountable final answer and reduce elapsed time for independent extraction tasks that do not share mutable state. Which plan should be approved to satisfy both checks?
- Perform both steps: (1) Encode a deterministic workflow with explicit transitions between stages; (2) enforce explicit completion checks and a maximum iteration budget.
- Perform both steps: (1) Use a supervisor agent to delegate bounded tasks and synthesize results; (2) run independent branches concurrently and merge their outputs afterward.
- Perform both steps: (1) Encode a deterministic workflow with explicit transitions between stages; (2) run independent branches concurrently and merge their outputs afterward.
- Perform both steps: (1) Use a supervisor agent to delegate bounded tasks and synthesize results; (2) enforce explicit completion checks and a maximum iteration budget.
Explanation
Correct Answer
B:The first change enables the team to coordinate several specialized investigations and assemble one accountable final answer because a supervisor owns the overall objective and integrates subagent findings; each delegate should have a narrow mandate and return a constrained summary; the second addresses the separate need to reduce elapsed time for independent extraction tasks that do not share mutable state because parallel orchestration shortens the critical path when calls have no dependencies, while the merger checks completeness before proceeding.
Incorrect Answers
A:Both steps target different results—run the same retrieval, validation, and publishing stages in an approved order for every request and stop iterative tool use when a success condition or resource budget is reached—because known control flow belongs in application code, allowing each stage to be tested and audited without asking a model to improvise the sequence and bounded loops avoid unending tool calls and uncontrolled spend while making incomplete work visible instead of silently continuing; the required capabilities are instead to coordinate several specialized investigations and assemble one accountable final answer and to reduce elapsed time for independent extraction tasks that do not share mutable state.
C:The second step meets the requirement to reduce elapsed time for independent extraction tasks that do not share mutable state because parallel orchestration shortens the critical path when calls have no dependencies, while the merger checks completeness before proceeding; the first step would only help to run the same retrieval, validation, and publishing stages in an approved order for every request because known control flow belongs in application code, allowing each stage to be tested and audited without asking a model to improvise the sequence, leaving the separate requirement to coordinate several specialized investigations and assemble one accountable final answer unmet.
D:The first step is appropriate for the requirement to coordinate several specialized investigations and assemble one accountable final answer because a supervisor owns the overall objective and integrates subagent findings; each delegate should have a narrow mandate and return a constrained summary; the other step instead aims to stop iterative tool use when a success condition or resource budget is reached because bounded loops avoid unending tool calls and uncontrolled spend while making incomplete work visible instead of silently continuing, which does not meet the second requirement to reduce elapsed time for independent extraction tasks that do not share mutable state.
Question 519
For a customer-facing AI feature, a design is acceptable only if it can ensure refunds above a policy limit receive human sign-off while also being able to need schema-conformant JSON for a downstream parser at generation time. Which paired implementation is sufficient?
- Perform both steps: (1) Add application-owned approval logic around the refund tool; (2) configure supported structured outputs with a json schema.
- Perform both steps: (1) Delimit retrieved text as untrusted source material; (2) set the appropriate application/json content type and serialize a valid payload.
- Perform both steps: (1) Delimit retrieved text as untrusted source material; (2) configure supported structured outputs with a json schema.
- Perform both steps: (1) Add application-owned approval logic around the refund tool; (2) set the appropriate application/json content type and serialize a valid payload.
Explanation
Correct Answer
A:The first change enables the team to ensure refunds above a policy limit receive human sign-off because decisions that create financial side effects require deterministic authorization checks beyond a model's recommendation; the second addresses the separate need to need schema-conformant JSON for a downstream parser at generation time because constrained JSON generation targets machine-parseable fields and types; free-form prompting alone does not establish the same guarantee.
Incorrect Answers
B:Both steps target different results—prevent a retrieved article's embedded commands from being treated as developer instructions and have the server parse a JSON request body predictably—because explicit trust boundaries help the model interpret documents as evidence rather than operational orders, and runtime controls must enforce privileged actions and a JSON body requires both correct serialization and matching HTTP metadata; raw language-model text cannot replace a valid request object; the required capabilities are instead to ensure refunds above a policy limit receive human sign-off and to need schema-conformant JSON for a downstream parser at generation time.
C:The second step meets the requirement to need schema-conformant JSON for a downstream parser at generation time because constrained JSON generation targets machine-parseable fields and types; free-form prompting alone does not establish the same guarantee; the first step would only help to prevent a retrieved article's embedded commands from being treated as developer instructions because explicit trust boundaries help the model interpret documents as evidence rather than operational orders, and runtime controls must enforce privileged actions, leaving the separate requirement to ensure refunds above a policy limit receive human sign-off unmet.
D:The first step is appropriate for the requirement to ensure refunds above a policy limit receive human sign-off because decisions that create financial side effects require deterministic authorization checks beyond a model's recommendation; the other step instead aims to have the server parse a JSON request body predictably because a JSON body requires both correct serialization and matching HTTP metadata; raw language-model text cannot replace a valid request object, which does not meet the second requirement to need schema-conformant JSON for a downstream parser at generation time.
Question 520
A team maintaining an enterprise MCP client has two different engineering requirements: (1) reject a date or identifier that violates business constraints before a service call, and (2) reduce context bloat caused by a tool returning a thousand irrelevant fields. Which TWO independent changes directly meet them? Choose TWO.
- Create a Skill containing the procedure and supporting instructions.
- Validate model-supplied tool arguments against schema and domain rules.
- Give each tool a timeout and controlled retry or failure result.
- Implement a narrowly scoped read-only tool handler.
- Return a concise task-relevant projection from the tool handler.
Explanation
Correct Answers
B:To reject a date or identifier that violates business constraints before a service call as part of a design that must also reduce context bloat caused by a tool returning a thousand irrelevant fields, even syntactically correct generated inputs can violate account scope or allowed values, so execution must validate both levels; this is why the action is one of the two required selections.
E:To reduce context bloat caused by a tool returning a thousand irrelevant fields as part of a design that must also reject a date or identifier that violates business constraints before a service call, tool output is subsequently consumed as model context; minimizing it reduces both token cost and distractions; this is why the action is one of the two required selections.
Incorrect Answers
A:This alternative serves the separate goal to make a repeatable multi-step code-review methodology available to an agent because a Skill teaches how to perform a task, whereas a tool is the executable capability that performs an external operation; the required outcomes are instead to reject a date or identifier that violates business constraints before a service call and to reduce context bloat caused by a tool returning a thousand irrelevant fields.
C:This alternative serves the separate goal to stop a slow external system from trapping an agent loop indefinitely because tool execution is a separate latency source; bounded calls prevent one dependency from exhausting the entire agent budget; the required outcomes are instead to reject a date or identifier that violates business constraints before a service call and to reduce context bloat caused by a tool returning a thousand irrelevant fields.
D:This alternative serves the separate goal to expose a diagnostic capability without allowing the model to modify records because a restricted read interface supports observation while preventing data mutation, unlike giving a diagnostic agent a general administrative API; the required outcomes are instead to reject a date or identifier that violates business constraints before a service call and to reduce context bloat caused by a tool returning a thousand irrelevant fields.