Question 200
A planned upgrade to a safety-sensitive Claude application has two requirements that must both pass: stop a coding agent from running commands that match a prohibited destructive pattern; reduce exposure after an API key is suspected compromised. Identify the two direct implementations. Choose TWO.
- Apply an input-side policy gate and route flagged cases to a safe path.
- Check resource-level authorization for every sensitive tool request.
- Revoke and rotate the affected credential and update dependent services.
- Enforce a pre-tool hook that denies the disallowed call.
- Use a server-side workload identity or scoped service credential.
Explanation
Correct Answers
C:To reduce exposure after an API key is suspected compromised as part of a design that must also stop a coding agent from running commands that match a prohibited destructive pattern, rotation restores credential control only when the old secret is disabled and consumers move to the replacement safely; this is why the action is one of the two required selections.
D:To stop a coding agent from running commands that match a prohibited destructive pattern as part of a design that must also reduce exposure after an API key is suspected compromised, a deterministic hook checks the proposed command before execution, providing a stronger control than asking the agent to self-police; this is why the action is one of the two required selections.
Incorrect Answers
A:This alternative serves the separate goal to identify prohibited requests before an automated agent invokes business tools because early filtering reduces inappropriate work, but later checks remain necessary because tool outputs can also introduce risks; the required outcomes are instead to stop a coding agent from running commands that match a prohibited destructive pattern and to reduce exposure after an API key is suspected compromised.
B:This alternative serves the separate goal to stop an authenticated user from reading another tenant's records via a tool because proving identity does not grant access to every object; the backend must enforce tenant and resource permissions separately; the required outcomes are instead to stop a coding agent from running commands that match a prohibited destructive pattern and to reduce exposure after an API key is suspected compromised.
E:This alternative serves the separate goal to stop frontend users from receiving unrestricted backend tool credentials because service identities tie privileges to a runtime workload and avoid exposing long-lived secrets to browsers or untrusted users; the required outcomes are instead to stop a coding agent from running commands that match a prohibited destructive pattern and to reduce exposure after an API key is suspected compromised.
Question 201
An architecture decision for a Claude API workload benchmark must support account for different plausible answers from the same input on repeated runs and run a straightforward well-specified classification without consuming examples. Which option gives a complete rather than partial solution?
- Perform both steps: (1) Model the response as conditioned token generation; (2) start with a clear instruction and no demonstrations.
- Perform both steps: (1) Model the response as conditioned token generation; (2) measure or count input and output tokens with the supported tokenizer or api.
- Perform both steps: (1) Design for stochastic output and evaluate behavior over representative repeats; (2) start with a clear instruction and no demonstrations.
- Perform both steps: (1) Design for stochastic output and evaluate behavior over representative repeats; (2) measure or count input and output tokens with the supported tokenizer or api.
Explanation
Correct Answer
C:The first change enables the team to account for different plausible answers from the same input on repeated runs because sampling and model execution can produce variation; important correctness must be checked rather than assumed identical on every call; the second addresses the separate need to run a straightforward well-specified classification without consuming examples because zero-shot prompting relies on direct task instructions and can save context when the task requires little format teaching.
Incorrect Answers
A:The second step meets the requirement to run a straightforward well-specified classification without consuming examples because zero-shot prompting relies on direct task instructions and can save context when the task requires little format teaching; 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 account for different plausible answers from the same input on repeated runs unmet.
B:Both steps target different results—explain why language-model output is generated step by step rather than fetched from a fixed response table and estimate model usage from encoded text rather than counting words—because autoregressive decoding selects subsequent tokens based on supplied context and prior output, which creates flexible but non-guaranteed wording and billing and context consumption track tokens, which do not map one-to-one to English words or characters; the required capabilities are instead to account for different plausible answers from the same input on repeated runs and to run a straightforward well-specified classification without consuming examples.
D:The first step is appropriate for the requirement to account for different plausible answers from the same input on repeated runs because sampling and model execution can produce variation; important correctness must be checked rather than assumed identical on every call; the other step instead aims 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, which does not meet the second requirement to run a straightforward well-specified classification without consuming examples.
Question 202
A design review for a high-volume Messages API gateway asks for support to trace exactly which prompt template was deployed with a code change, together with a way to debug a failed request without exposing the user's confidential document. Select the two appropriate changes. Choose TWO.
- Authenticate incoming callbacks and validate their signatures where supported.
- Log status, timing, and correlation IDs with sensitive fields redacted.
- Reuse an appropriately configured HTTP client or SDK connection pool.
- Version prompt files alongside relevant application code.
- Use the user's own Claude Code settings file.
Explanation
Correct Answers
B:To debug a failed request without exposing the user's confidential document as part of a design that must also trace exactly which prompt template was deployed with a code change, operational records can identify a failure path without retaining raw secrets or protected prompt content; this is why the action is one of the two required selections.
D:To trace exactly which prompt template was deployed with a code change as part of a design that must also debug a failed request without exposing the user's confidential document, prompts change software behavior and should be reviewed and diffed like code, rather than edited anonymously in production; this is why the action is one of the two required selections.
Incorrect Answers
A:This alternative serves the separate goal to trust a callback only after verifying who actually sent it because network arrival alone does not prove a webhook's origin; authenticity checks prevent forged events from changing agent state; the required outcomes are instead to trace exactly which prompt template was deployed with a code change and to debug a failed request without exposing the user's confidential document.
C:This alternative serves the separate goal to reduce repeated TLS setup overhead for high-volume application traffic because connection reuse lowers transport overhead but still requires limits and health handling for long-lived clients; the required outcomes are instead to trace exactly which prompt template was deployed with a code change and to debug a failed request without exposing the user's confidential document.
E:This alternative serves the separate goal to customize one developer's default Claude Code behavior across local projects because user settings are scoped to one person, so they should not be relied on for enforcing team-wide policy; the required outcomes are instead to trace exactly which prompt template was deployed with a code change and to debug a failed request without exposing the user's confidential document.
Question 203
While planning a tool-using incident triage agent, the product owner specifies that it must use a Claude platform-provided capability without operating a local handler. Which solution fulfills the requirement?
- Configure the appropriate Anthropic-executed server tool.
- Use the Claude Agent SDK for the agent's execution harness.
- Deploy the agent runner in the organization's own environment.
- Implement an application-owned Messages API agent loop.
Explanation
Correct Answer
A:To use a Claude platform-provided capability without operating a local handler in this tool-using incident triage agent, server tools execute on the provider side and have distinct authorization and runtime semantics from client tools the application dispatches; the proposed implementation directly supplies that capability.
Incorrect Answers
B:The proposed alternative helps to build a Claude-powered coding agent with an established runtime for tool orchestration because the SDK provides structured primitives for agent sessions and tool execution, reducing the custom loop infrastructure the application must maintain; that mechanism does not meet the different requirement to use a Claude platform-provided capability without operating a local handler.
C:The proposed alternative helps to keep orchestration runtime and internal tool handlers inside company-managed infrastructure because self-hosting puts tool execution, networking, and operational controls under the application's responsibility, subject to its own deployment constraints; that mechanism does not meet the different requirement to use a Claude platform-provided capability without operating a local handler.
D:The proposed alternative helps to require precise control over message assembly, tool dispatch, and stopping behavior because a custom loop explicitly owns each model call, tool result, state update, and budget decision, trading flexibility for more engineering responsibility; that mechanism does not meet the different requirement to use a Claude platform-provided capability without operating a local handler.
Question 204
In a Claude inference cost review, a release requirement is to detect a looping agent that suddenly multiplies token consumption. Which implementation addresses that specific requirement?
- Place caching around stable ordered content before dynamic material.
- Alert on unusual per-job token and tool-call growth.
- Iterate the SDK or HTTP event stream until its terminal event.
- Use asynchronous client methods and limit in-flight calls.
Explanation
Correct Answer
B:To detect a looping agent that suddenly multiplies token consumption in this Claude inference cost review, monitoring consumption across an entire agent run catches runaway workflows that per-request logs may not reveal; the proposed implementation directly supplies that capability.
Incorrect Answers
A:The proposed alternative helps to prevent constantly changing user text from invalidating a reusable prompt prefix because cache matching operates over the stable prefix, so volatile material before the breakpoint undermines reuse; that mechanism does not meet the different requirement to detect a looping agent that suddenly multiplies token consumption.
C:The proposed alternative helps to process partial model output as it is delivered over a streaming connection because streaming clients must handle both content deltas and final status rather than assuming the first chunk is the complete response; that mechanism does not meet the different requirement to detect a looping agent that suddenly multiplies token consumption.
D:The proposed alternative helps to keep a server responsive while several independent model calls wait on network I/O because event-loop based I/O overlaps waits, but rate and resource bounds are still necessary to avoid overload; that mechanism does not meet the different requirement to detect a looping agent that suddenly multiplies token consumption.
Question 205
Two approval criteria are still open for a structured output service: the team must require that a generated report have a fixed set of sections and apply the same compliance restrictions across many different end-user requests. Which two implementation actions satisfy them? Choose TWO.
- Put stable application constraints in the system instruction field.
- Name each mandatory field or section and provide formatting constraints.
- Specify the forbidden content alongside an enforceable validation check.
- Validate parsed output against the application's declared schema.
- Revise one prompt component at a time against a regression evaluation set.
Explanation
Correct Answers
A:To apply the same compliance restrictions across many different end-user requests as part of a design that must also require that a generated report have a fixed set of sections, high-priority application instructions should not be delivered as ordinary user content where they compete with user questions and document text; this is why the action is one of the two required selections.
B:To require that a generated report have a fixed set of sections as part of a design that must also apply the same compliance restrictions across many different end-user requests, concrete desired structure is easier to verify than a vague instruction to be professional or comprehensive; this is why the action is one of the two required selections.
Incorrect Answers
C:This alternative serves the separate goal to avoid a prohibited output element such as confidential identifiers because negative prompt guidance signals the intended boundary, but a deterministic validator is needed where violating it has consequences; the required outcomes are instead to require that a generated report have a fixed set of sections and to apply the same compliance restrictions across many different end-user requests.
D:This alternative serves the separate goal to reject model output missing a field that a downstream service requires because parsing proves a string is legal JSON, but schema validation checks required keys, types, and structural constraints; the required outcomes are instead to require that a generated report have a fixed set of sections and to apply the same compliance restrictions across many different end-user requests.
E:This alternative serves the separate goal to improve a prompt while preserving performance on previously solved cases because controlled prompt iteration distinguishes genuine quality gains from regressions or overfitting to one attractive example; the required outcomes are instead to require that a generated report have a fixed set of sections and to apply the same compliance restrictions across many different end-user requests.
Question 206
A test for a safety-sensitive Claude application fails because it cannot require a person to approve an irreversible administrative change. Which change fixes the identified failure?
- Pause execution for explicit approval at the high-impact action boundary.
- Revoke and rotate the affected credential and update dependent services.
- Redact secrets and protect log access with retention controls.
- Combine prompt boundaries, scoped credentials, validation, and approval gates.
Explanation
Correct Answer
A:To require a person to approve an irreversible administrative change in this safety-sensitive Claude application, human approval is a policy-enforced transition, not a note written in the model's plan or a message displayed afterward; the proposed implementation directly supplies that capability.
Incorrect Answers
B:The proposed alternative helps to reduce exposure after an API key is suspected compromised because rotation restores credential control only when the old secret is disabled and consumers move to the replacement safely; that mechanism does not meet the different requirement to require a person to approve an irreversible administrative change.
C:The proposed alternative helps to investigate user actions without leaking API keys into operational records because logs support accountability only when access and retention are restricted; copying all prompt content into logs creates another exposure surface; that mechanism does not meet the different requirement to require a person to approve an irreversible administrative change.
D:The proposed alternative helps 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; that mechanism does not meet the different requirement to require a person to approve an irreversible administrative change.
Question 207
A release check on an autonomous repository maintenance job found two unsatisfied conditions: the service must stop old agent histories from burying the information needed for the current step, and it must prevent a research assistant from unexpectedly making production changes. Which complete engineering plan satisfies both?
- Perform both steps: (1) Retrieve only task-relevant memories into the active context; (2) return a separate tool_result for every corresponding tool_use identifier.
- Perform both steps: (1) Retrieve only task-relevant memories into the active context; (2) give the research subagent read-only tools and a narrow instruction contract.
- Perform both steps: (1) Prefer a simple API call over a general-purpose multi-agent framework; (2) give the research subagent read-only tools and a narrow instruction contract.
- Perform both steps: (1) Prefer a simple API call over a general-purpose multi-agent framework; (2) return a separate tool_result for every corresponding tool_use identifier.
Explanation
Correct Answer
B:The first change enables the team to stop old agent histories from burying the information needed for the current step because selective retrieval improves signal-to-noise and keeps token use bounded while retaining access to older information when required; the second addresses the separate need to prevent a research assistant from unexpectedly making production changes because capability isolation prevents a specialist from acting outside its delegated role even if it encounters persuasive untrusted instructions.
Incorrect Answers
A:The first step is appropriate for the requirement to stop old agent histories from burying the information needed for the current step because selective retrieval improves signal-to-noise and keeps token use bounded while retaining access to older information when required; the other step instead aims to handle several independent tool_use blocks in one assistant message because parallel tool execution requires complete, correctly linked results; dropping one identifier leaves the model's outstanding request unresolved, which does not meet the second requirement to prevent a research assistant from unexpectedly making production changes.
C:The second step meets the requirement to prevent a research assistant from unexpectedly making production changes because capability isolation prevents a specialist from acting outside its delegated role even if it encounters persuasive untrusted instructions; the first step would only help 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, leaving the separate requirement to stop old agent histories from burying the information needed for the current step unmet.
D:Both steps target different results—avoid adding orchestration infrastructure to a single request-and-response feature and handle several independent tool_use blocks in one assistant message—because frameworks are useful for genuinely branching, long-lived execution, but increase complexity when no orchestration problem exists and parallel tool execution requires complete, correctly linked results; dropping one identifier leaves the model's outstanding request unresolved; the required capabilities are instead to stop old agent histories from burying the information needed for the current step and to prevent a research assistant from unexpectedly making production changes.
Question 208
Two acceptance tests block a production inference application: one requires the ability to make a direct REST request with a declared supported API contract; the other requires the ability to send a model inference request with the operation expected by the REST endpoint. Which plan clears both tests?
- Perform both steps: (1) Version prompt files alongside relevant application code; (2) use the api's documented post request with a json body.
- Perform both steps: (1) Supply the required API version header; (2) validate file type, size, and allowed content conversion before api submission.
- Perform both steps: (1) Version prompt files alongside relevant application code; (2) validate file type, size, and allowed content conversion before api submission.
- Perform both steps: (1) Supply the required API version header; (2) use the api's documented post request with a json body.
Explanation
Correct Answer
D:The first change enables the team to make a direct REST request with a declared supported API contract because the version header selects the protocol contract supported by the endpoint, preventing accidental dependence on unspecified behavior; the second addresses the separate need to send a model inference request with the operation expected by the REST endpoint because a state-changing inference operation is submitted as an HTTP request with an explicit method and payload, not inferred from a browser URL alone.
Incorrect Answers
A:The second step meets the requirement to send a model inference request with the operation expected by the REST endpoint because a state-changing inference operation is submitted as an HTTP request with an explicit method and payload, not inferred from a browser URL alone; the first step would only help to trace exactly which prompt template was deployed with a code change because prompts change software behavior and should be reviewed and diffed like code, rather than edited anonymously in production, leaving the separate requirement to make a direct REST request with a declared supported API contract unmet.
B:The first step is appropriate for the requirement to make a direct REST request with a declared supported API contract because the version header selects the protocol contract supported by the endpoint, preventing accidental dependence on unspecified behavior; the other step instead aims to prevent a document-upload endpoint from passing unsupported bytes to the model because client filenames are not enough to establish content validity; ingestion must enforce protocol and product constraints, which does not meet the second requirement to send a model inference request with the operation expected by the REST endpoint.
C:Both steps target different results—trace exactly which prompt template was deployed with a code change and prevent a document-upload endpoint from passing unsupported bytes to the model—because prompts change software behavior and should be reviewed and diffed like code, rather than edited anonymously in production and client filenames are not enough to establish content validity; ingestion must enforce protocol and product constraints; the required capabilities are instead to make a direct REST request with a declared supported API contract and to send a model inference request with the operation expected by the REST endpoint.
Question 209
While planning a long-context summarization workflow, the product owner specifies that it must provide a document's necessary meaning without disclosing irrelevant personal identifiers. Which solution fulfills the requirement?
- State the task, audience, scope, and required deliverable explicitly.
- Rank and select relevant passages before building the model context.
- Place essential constraints prominently and structure large evidence clearly.
- Minimize or redact sensitive fields before constructing context.
Explanation
Correct Answer
D:To provide a document's necessary meaning without disclosing irrelevant personal identifiers in this long-context summarization workflow, data minimization reduces the sensitive information transmitted for inference while retaining the facts the task legitimately requires; the proposed implementation directly supplies that capability.
Incorrect Answers
A:The proposed alternative helps to stop a model from alternating unpredictably between summary and extraction because specific task wording narrows reasonable interpretations and gives validation rules a clear target without relying on implied intent; that mechanism does not meet the different requirement to provide a document's necessary meaning without disclosing irrelevant personal identifiers.
B:The proposed alternative helps to keep only evidence that matters to the question when retrieval returns hundreds of documents because relevance filtering reduces distracting information and token usage; dumping every retrieved page can degrade grounding and exceed limits; that mechanism does not meet the different requirement to provide a document's necessary meaning without disclosing irrelevant personal identifiers.
C:The proposed alternative helps to reduce the chance that critical instructions become buried in a huge prompt because ordering, headings, and explicit references help the model distinguish governing rules from supporting material in large requests; that mechanism does not meet the different requirement to provide a document's necessary meaning without disclosing irrelevant personal identifiers.
Question 210
The owner of an internal agent-extension rollout needs to reduce context bloat caused by a tool returning a thousand irrelevant fields as well as process several requested independent tools without mixing their outputs. Which two changes belong in the minimum remediation plan? Choose TWO.
- Publish a properly described MCP tool and its input contract.
- Correlate each tool result with its matching tool_use_id.
- Return an explicit tool error indication and an informative safe message.
- Return a concise task-relevant projection from the tool handler.
- Configure a supported remote HTTP transport with authentication and deployment controls.
Explanation
Correct Answers
B:To process several requested independent tools without mixing their outputs as part of a design that must also reduce context bloat caused by a tool returning a thousand irrelevant fields, multi-tool responses can interleave operations, so identifier-based correlation is essential for safe continuation; this is why the action is one of the two required selections.
D:To reduce context bloat caused by a tool returning a thousand irrelevant fields as part of a design that must also process several requested independent tools without mixing their outputs, 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 let an MCP client invoke a business capability through a standard server interface because MCP tools represent callable operations with typed parameters; the server executes authorized behavior rather than only providing reference text; the required outcomes are instead to reduce context bloat caused by a tool returning a thousand irrelevant fields and to process several requested independent tools without mixing their outputs.
C:This alternative serves the separate goal to let Claude distinguish a failed database lookup from an empty successful result because an error result preserves failure semantics, preventing the model from inventing a no-record finding for a call that never succeeded; the required outcomes are instead to reduce context bloat caused by a tool returning a thousand irrelevant fields and to process several requested independent tools without mixing their outputs.
E:This alternative serves the separate goal to expose an MCP server to clients across a network boundary because remote connectivity needs network routing, identity, and authorization; a local stdio subprocess is not a remotely reachable service; the required outcomes are instead to reduce context bloat caused by a tool returning a thousand irrelevant fields and to process several requested independent tools without mixing their outputs.