Question 680
Two acceptance tests block a production inference application: one requires the ability to preserve behavior while replacing a fragile parser in an AI integration; the other requires the ability to detect a changed tool response field before it breaks downstream code. Which plan clears both tests?
- Perform both steps: (1) Use the API's documented POST request with a JSON body; (2) log status, timing, and correlation ids with sensitive fields redacted.
- Perform both steps: (1) Refactor in bounded steps under existing tests; (2) log status, timing, and correlation ids with sensitive fields redacted.
- Perform both steps: (1) Refactor in bounded steps under existing tests; (2) add automated consumer-provider contract tests.
- Perform both steps: (1) Use the API's documented POST request with a JSON body; (2) add automated consumer-provider contract tests.
Explanation
Correct Answer
C:The first change enables the team to preserve behavior while replacing a fragile parser in an AI integration because small changes constrain regression scope; a wholesale rewrite makes it harder to isolate whether model or parsing behavior changed; the second addresses the separate need to detect a changed tool response field before it breaks downstream code because schema contracts make integration changes visible and prevent a renamed field from silently corrupting later model context.
Incorrect Answers
A:Both steps target different results—send a model inference request with the operation expected by the REST endpoint and debug a failed request without exposing the user's confidential document—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 and operational records can identify a failure path without retaining raw secrets or protected prompt content; the required capabilities are instead to preserve behavior while replacing a fragile parser in an AI integration and to detect a changed tool response field before it breaks downstream code.
B:The first step is appropriate for the requirement to preserve behavior while replacing a fragile parser in an AI integration because small changes constrain regression scope; a wholesale rewrite makes it harder to isolate whether model or parsing behavior changed; the other step instead aims to debug a failed request without exposing the user's confidential document because operational records can identify a failure path without retaining raw secrets or protected prompt content, which does not meet the second requirement to detect a changed tool response field before it breaks downstream code.
D:The second step meets the requirement to detect a changed tool response field before it breaks downstream code because schema contracts make integration changes visible and prevent a renamed field from silently corrupting later model context; the first step would only help 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, leaving the separate requirement to preserve behavior while replacing a fragile parser in an AI integration unmet.
Question 681
A team maintaining an enterprise agent deployment has two different engineering requirements: (1) stop a model-requested webhook from sending sensitive case text to an unapproved host, and (2) investigate user actions without leaking API keys into operational records. Which TWO independent changes directly meet them? Choose TWO.
- Treat retrieved content as untrusted and prevent its instructions from controlling tools.
- Grant each runtime tool only the operations and resources required for its task.
- Use layered policy enforcement and restrict powerful tools independently of chat text.
- Redact secrets and protect log access with retention controls.
- Enforce outbound destination allowlists at the tool execution layer.
Explanation
Correct Answers
D:To investigate user actions without leaking API keys into operational records as part of a design that must also stop a model-requested webhook from sending sensitive case text to an unapproved host, logs support accountability only when access and retention are restricted; copying all prompt content into logs creates another exposure surface; this is why the action is one of the two required selections.
E:To stop a model-requested webhook from sending sensitive case text to an unapproved host as part of a design that must also investigate user actions without leaking API keys into operational records, network restrictions prevent data egress even if injected instructions convince a model to request the call; this is why the action is one of the two required selections.
Incorrect Answers
A:This alternative serves the separate goal to resist instructions embedded in a retrieved support ticket that demand secret disclosure because source material is data rather than authority; enforcing tool permissions and separating roles limits an attacker's ability to redirect execution; the required outcomes are instead to stop a model-requested webhook from sending sensitive case text to an unapproved host and to investigate user actions without leaking API keys into operational records.
B:This alternative serves the separate goal to limit the consequences of a mistaken tool call in a production support system because least-privilege permissions shrink the blast radius of model errors and compromised prompts, especially for mutating actions; the required outcomes are instead to stop a model-requested webhook from sending sensitive case text to an unapproved host and to investigate user actions without leaking API keys into operational records.
C:This alternative serves the separate goal to prevent an adversarial user from bypassing safety policy by roleplay instructions because adversarial prompting can target instruction hierarchy, so product controls and tool authorization must remain effective when wording is manipulated; the required outcomes are instead to stop a model-requested webhook from sending sensitive case text to an unapproved host and to investigate user actions without leaking API keys into operational records.
Question 682
The technical owner of an internal coding-assistant workflow has to use a Claude platform-provided capability without operating a local handler. A separate requirement is to build a Claude-powered coding agent with an established runtime for tool orchestration. Which two-part remediation covers both?
- Perform both steps: (1) Configure the appropriate Anthropic-executed server tool; (2) write selected durable facts to a controlled external memory store.
- Perform both steps: (1) Apply limits for steps, elapsed time, and token or tool expenditure; (2) write selected durable facts to a controlled external memory store.
- Perform both steps: (1) Apply limits for steps, elapsed time, and token or tool expenditure; (2) use the claude agent sdk for the agent's execution harness.
- Perform both steps: (1) Configure the appropriate Anthropic-executed server tool; (2) use the claude agent sdk for the agent's execution harness.
Explanation
Correct Answer
D:The first change enables the team to use a Claude platform-provided capability without operating a local handler because server tools execute on the provider side and have distinct authorization and runtime semantics from client tools the application dispatches; the second addresses the separate need 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.
Incorrect Answers
A:The first step is appropriate for the requirement to use a Claude platform-provided capability without operating a local handler because server tools execute on the provider side and have distinct authorization and runtime semantics from client tools the application dispatches; the other step instead aims to preserve durable user-approved facts across agent sessions outside the context window because external memory allows selective retrieval in later sessions, unlike temporary model context that disappears when calls end, which does not meet the second requirement to build a Claude-powered coding agent with an established runtime for tool orchestration.
B:Both steps target different results—impose predictable maximum cost and wall-clock exposure on autonomous execution and preserve durable user-approved facts across agent sessions outside the context window—because runtime budgets are external constraints on agent activity and must be enforced by the host, not inferred from a cooperative natural-language instruction and external memory allows selective retrieval in later sessions, unlike temporary model context that disappears when calls end; the required capabilities are instead to use a Claude platform-provided capability without operating a local handler and to build a Claude-powered coding agent with an established runtime for tool orchestration.
C:The second step meets the requirement 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; the first step would only help to impose predictable maximum cost and wall-clock exposure on autonomous execution because runtime budgets are external constraints on agent activity and must be enforced by the host, not inferred from a cooperative natural-language instruction, leaving the separate requirement to use a Claude platform-provided capability without operating a local handler unmet.
Question 683
A production readiness review of a hosted document-processing service finds that the team cannot yet prevent a retrieved article's embedded commands from being treated as developer instructions or prevent chat transcripts from different accounts entering the same Claude request. Which paired changes directly remedy the two gaps?
- Perform both steps: (1) Route Claude requests through a trusted backend service; (2) store and retrieve message history by authenticated tenant and conversation id.
- Perform both steps: (1) Delimit retrieved text as untrusted source material; (2) configure supported structured outputs with a json schema.
- Perform both steps: (1) Delimit retrieved text as untrusted source material; (2) store and retrieve message history by authenticated tenant and conversation id.
- Perform both steps: (1) Route Claude requests through a trusted backend service; (2) configure supported structured outputs with a json schema.
Explanation
Correct Answer
C:The first change enables the team 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; the second addresses the separate need to prevent chat transcripts from different accounts entering the same Claude request because session hygiene requires server-enforced isolation; model instructions cannot repair a history assembled from the wrong user.
Incorrect Answers
A:The second step meets the requirement to prevent chat transcripts from different accounts entering the same Claude request because session hygiene requires server-enforced isolation; model instructions cannot repair a history assembled from the wrong user; the first step would only help to keep a browser application from exposing the secret used for model calls because a backend controls credentials and access policy while the UI sends user input through an authenticated application endpoint, leaving the separate requirement to prevent a retrieved article's embedded commands from being treated as developer instructions unmet.
B:The first step is appropriate for the requirement 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; the other step instead aims 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, which does not meet the second requirement to prevent chat transcripts from different accounts entering the same Claude request.
D:Both steps target different results—keep a browser application from exposing the secret used for model calls and need schema-conformant JSON for a downstream parser at generation time—because a backend controls credentials and access policy while the UI sends user input through an authenticated application endpoint and constrained JSON generation targets machine-parseable fields and types; free-form prompting alone does not establish the same guarantee; the required capabilities are instead to prevent a retrieved article's embedded commands from being treated as developer instructions and to prevent chat transcripts from different accounts entering the same Claude request.
Question 684
For a Claude inference cost review, a design is acceptable only if it can avoid overlooking slow tail requests hidden by a good average while also being able to detect a looping agent that suddenly multiplies token consumption. Which paired implementation is sufficient?
- Perform both steps: (1) Measure cache-write overhead against subsequent cache-read savings; (2) propagate cancellation to the active network request when supported.
- Perform both steps: (1) Measure cache-write overhead against subsequent cache-read savings; (2) alert on unusual per-job token and tool-call growth.
- Perform both steps: (1) Report p50, p95, and p99 latency by model and operation; (2) propagate cancellation to the active network request when supported.
- Perform both steps: (1) Report p50, p95, and p99 latency by model and operation; (2) alert on unusual per-job token and tool-call growth.
Explanation
Correct Answer
D:The first change enables the team to avoid overlooking slow tail requests hidden by a good average because percentiles expose user experience at the tail and can reveal bottlenecks that a mean response time conceals; the second addresses the separate need to detect a looping agent that suddenly multiplies token consumption because monitoring consumption across an entire agent run catches runaway workflows that per-request logs may not reveal.
Incorrect Answers
A:Both steps target different results—reduce repeated cost for a stable lengthy policy prefix reused many times and stop a no-longer-needed inference when the user's request is cancelled—because cache writes may cost extra while hits are cheaper; the reuse frequency determines whether the optimization saves money and cancelling work can free local resources and avoid processing obsolete results, though already-incurred provider work may still be billed; the required capabilities are instead to avoid overlooking slow tail requests hidden by a good average and to detect a looping agent that suddenly multiplies token consumption.
B:The second step meets the requirement to detect a looping agent that suddenly multiplies token consumption because monitoring consumption across an entire agent run catches runaway workflows that per-request logs may not reveal; the first step would only help 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, leaving the separate requirement to avoid overlooking slow tail requests hidden by a good average unmet.
C:The first step is appropriate for the requirement to avoid overlooking slow tail requests hidden by a good average because percentiles expose user experience at the tail and can reveal bottlenecks that a mean response time conceals; the other step instead aims to stop a no-longer-needed inference when the user's request is cancelled because cancelling work can free local resources and avoid processing obsolete results, though already-incurred provider work may still be billed, which does not meet the second requirement to detect a looping agent that suddenly multiplies token consumption.
Question 685
The team responsible for a shared tool gateway for Claude needs to remove an agent's access to an obsolete integration without rewriting all its task prompts. What is the most direct engineering change?
- Create a Skill containing the procedure and supporting instructions.
- Disable the capability at configuration or tool-registration level.
- Expose the manual as an MCP resource rather than a tool.
- Implement a client-side handler and return its result with the tool-use identifier.
Explanation
Correct Answer
B:To remove an agent's access to an obsolete integration without rewriting all its task prompts in this shared tool gateway for Claude, actual registered capabilities define what may execute; changing descriptive prompt text alone does not revoke a previously available tool; the proposed implementation directly supplies that capability.
Incorrect Answers
A:The proposed alternative helps 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; that mechanism does not meet the different requirement to remove an agent's access to an obsolete integration without rewriting all its task prompts.
C:The proposed alternative helps to supply a read-only policy manual to an MCP client without giving it an action endpoint because a resource is contextual data; representing a passive document as an imperative function adds unnecessary execution semantics; that mechanism does not meet the different requirement to remove an agent's access to an obsolete integration without rewriting all its task prompts.
D:The proposed alternative helps to connect a Claude tool request to an internal service whose logic the application owns because a declared client tool does not execute itself; application code runs it and feeds results back into the conversation; that mechanism does not meet the different requirement to remove an agent's access to an obsolete integration without rewriting all its task prompts.
Question 686
A team maintaining a Python Claude API backend has two different engineering requirements: (1) stop a well-formed JSON response from introducing an impossible customer status, and (2) need schema-conformant JSON for a downstream parser at generation time. Which TWO independent changes directly meet them? Choose TWO.
- Buffer structured data until it can be parsed and validated before UI commit.
- Validate generated fields against domain rules after parsing.
- Retain source identifiers with retrieved evidence and render supported citations.
- Configure supported structured outputs with a JSON Schema.
- Use the API's documented POST request with a JSON body.
Explanation
Correct Answers
B:To stop a well-formed JSON response from introducing an impossible customer status as part of a design that must also need schema-conformant JSON for a downstream parser at generation time, a syntactically valid object may still contain invalid business values; server-side semantics determine whether it is acceptable; this is why the action is one of the two required selections.
D:To need schema-conformant JSON for a downstream parser at generation time as part of a design that must also stop a well-formed JSON response from introducing an impossible customer status, constrained JSON generation targets machine-parseable fields and types; free-form prompting alone does not establish the same guarantee; this is why the action is one of the two required selections.
Incorrect Answers
A:This alternative serves the separate goal to display partial sentences without publishing malformed structured records because a text stream can arrive in fragments; consumers must distinguish presentation updates from complete machine-readable objects; the required outcomes are instead to stop a well-formed JSON response from introducing an impossible customer status and to need schema-conformant JSON for a downstream parser at generation time.
C:This alternative serves the separate goal to let a reviewer trace claims to the material actually consulted because traceability requires linking claims to accessible source passages; fabricated URLs or unlabeled snippets do not provide verifiable provenance; the required outcomes are instead to stop a well-formed JSON response from introducing an impossible customer status and to need schema-conformant JSON for a downstream parser at generation time.
E:This alternative serves the separate goal 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 required outcomes are instead to stop a well-formed JSON response from introducing an impossible customer status and to need schema-conformant JSON for a downstream parser at generation time.
Question 687
An architecture review of a shared Claude Code repository asks how to apply a coding convention only when Claude Code edits a particular path. Which action is the best fit?
- Add concise project guidance to the appropriate CLAUDE.md file.
- Initialize project guidance and review the resulting CLAUDE.md content.
- Create a project skill with the documented skill metadata and content.
- Place a path-targeted rule in the repository's rules configuration.
Explanation
Correct Answer
D:To apply a coding convention only when Claude Code edits a particular path in this shared Claude Code repository, scoped rules keep specialized instructions close to their applicable files, avoiding global instructions that conflict with other modules; the proposed implementation directly supplies that capability.
Incorrect Answers
A:The proposed alternative helps to give Claude Code repository-specific build and test instructions for future sessions because Claude Code reads persistent project instructions from its memory hierarchy; one-off chat messages do not reliably establish durable repository guidance; that mechanism does not meet the different requirement to apply a coding convention only when Claude Code edits a particular path.
B:The proposed alternative helps to introduce Claude Code to an unfamiliar codebase's conventions and commands because repository setup discovers useful project practices, but autogenerated guidance still needs a human check for accuracy and unnecessary detail; that mechanism does not meet the different requirement to apply a coding convention only when Claude Code edits a particular path.
C:The proposed alternative helps to make a reusable development procedure available to Claude Code with supporting files because a skill packages guidance or a workflow for reuse across relevant tasks; it is not identical to granting a runtime tool permission; that mechanism does not meet the different requirement to apply a coding convention only when Claude Code edits a particular path.
Question 688
A change request for a Claude integration debugging session combines two needs: distinguish a syntactically valid but low-quality answer from a transport outage, and find why a JSON record unexpectedly ends midway through a field. Which implementation pair addresses both, not just one?
- Perform both steps: (1) Read the 400 error details and validate the request payload; (2) inspect stop_reason and output limits before changing the parser.
- Perform both steps: (1) Compare versioned prompts and representative eval cases against model responses; (2) inspect the http 429 status and rate-limit metadata before retrying.
- Perform both steps: (1) Compare versioned prompts and representative eval cases against model responses; (2) inspect stop_reason and output limits before changing the parser.
- Perform both steps: (1) Read the 400 error details and validate the request payload; (2) inspect the http 429 status and rate-limit metadata before retrying.
Explanation
Correct Answer
C:The first change enables the team to distinguish a syntactically valid but low-quality answer from a transport outage because a successful HTTP call only proves the protocol worked; quality regressions require output-level evidence and controlled comparisons; the second addresses the separate need to find why a JSON record unexpectedly ends midway through a field because a max_tokens stop indicates incomplete generation; the parser may be working correctly against a response that was never finished.
Incorrect Answers
A:The second step meets the requirement to find why a JSON record unexpectedly ends midway through a field because a max_tokens stop indicates incomplete generation; the parser may be working correctly against a response that was never finished; the first step would only help to find a request schema error instead of repeatedly changing the system prompt because an invalid API field or incompatible parameter cannot be fixed by natural-language instructions within the malformed request, leaving the separate requirement to distinguish a syntactically valid but low-quality answer from a transport outage unmet.
B:The first step is appropriate for the requirement to distinguish a syntactically valid but low-quality answer from a transport outage because a successful HTTP call only proves the protocol worked; quality regressions require output-level evidence and controlled comparisons; the other step instead aims to distinguish an upstream throttling response from a malformed model answer because a rate limit is a request-layer capacity problem; the response needs controlled pacing rather than a prompt rewrite, which does not meet the second requirement to find why a JSON record unexpectedly ends midway through a field.
D:Both steps target different results—find a request schema error instead of repeatedly changing the system prompt and distinguish an upstream throttling response from a malformed model answer—because an invalid API field or incompatible parameter cannot be fixed by natural-language instructions within the malformed request and a rate limit is a request-layer capacity problem; the response needs controlled pacing rather than a prompt rewrite; the required capabilities are instead to distinguish a syntactically valid but low-quality answer from a transport outage and to find why a JSON record unexpectedly ends midway through a field.
Question 689
While planning a business workflow that calls Claude, the product owner specifies that it must expose external business tools only to workflows that actually need them. Which solution fulfills the requirement?
- Require structured request IDs, tool traces, and version metadata with privacy controls.
- Track enabled plugins and their compatible versions as release dependencies.
- Retain source identifiers with retrieved evidence and render supported citations.
- Enable required plugin or connector capabilities using least-privilege configuration.
Explanation
Correct Answer
D:To expose external business tools only to workflows that actually need them in this business workflow that calls Claude, tools expand the application's action surface; indiscriminate activation increases risk and can confuse model selection among similar functions; the proposed implementation directly supplies that capability.
Incorrect Answers
A:The proposed alternative helps to enable an administrator to determine how a particular AI decision was produced because traceability depends on consistent operational evidence, not just retaining the final generated answer or a screenshot; that mechanism does not meet the different requirement to expose external business tools only to workflows that actually need them.
B:The proposed alternative helps to avoid a plugin update unexpectedly changing tools available to a coding agent because plugin capability changes can affect tool names and permissions, so production configuration should be reproducible and reviewed; that mechanism does not meet the different requirement to expose external business tools only to workflows that actually need them.
C:The proposed alternative helps to let a reviewer trace claims to the material actually consulted because traceability requires linking claims to accessible source passages; fabricated URLs or unlabeled snippets do not provide verifiable provenance; that mechanism does not meet the different requirement to expose external business tools only to workflows that actually need them.
Question 690
The technical owner of an internal coding-assistant workflow has to prevent a research assistant from unexpectedly making production changes. A separate requirement is to recover a multi-hour agent workflow after an infrastructure interruption. Which two-part remediation covers both?
- Perform both steps: (1) Use isolated branch state and a controlled reconciliation step; (2) persist checkpoints for tool outputs and accepted decisions.
- Perform both steps: (1) Give the research subagent read-only tools and a narrow instruction contract; (2) write selected durable facts to a controlled external memory store.
- Perform both steps: (1) Use isolated branch state and a controlled reconciliation step; (2) write selected durable facts to a controlled external memory store.
- Perform both steps: (1) Give the research subagent read-only tools and a narrow instruction contract; (2) persist checkpoints for tool outputs and accepted decisions.
Explanation
Correct Answer
D:The first change enables the team 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 second addresses the separate need to recover a multi-hour agent workflow after an infrastructure interruption because durable checkpoints let the orchestrator resume from a known stage rather than rerunning costly or side-effecting actions.
Incorrect Answers
A:The second step meets the requirement to recover a multi-hour agent workflow after an infrastructure interruption because durable checkpoints let the orchestrator resume from a known stage rather than rerunning costly or side-effecting actions; the first step would only help 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, leaving the separate requirement to prevent a research assistant from unexpectedly making production changes unmet.
B:The first step is appropriate for 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 other step instead aims to preserve durable user-approved facts across agent sessions outside the context window because external memory allows selective retrieval in later sessions, unlike temporary model context that disappears when calls end, which does not meet the second requirement to recover a multi-hour agent workflow after an infrastructure interruption.
C:Both steps target different results—stop two parallel branches from overwriting the same mutable record and preserve durable user-approved facts across agent sessions outside the context window—because explicit merge rules are required when outputs converge; concurrent mutation without arbitration causes nondeterministic results and external memory allows selective retrieval in later sessions, unlike temporary model context that disappears when calls end; the required capabilities are instead to prevent a research assistant from unexpectedly making production changes and to recover a multi-hour agent workflow after an infrastructure interruption.