Question 50
A Claude application review asks the team to close two unrelated gaps without merging their execution paths. Service A must avoid assuming a Claude Code repository setting automatically affects a direct Messages API call; service B must avoid exposing raw production database credentials directly to an agent. Which paired remediation works?
- First component: Place durable behavior instructions in the system parameter. Second component: Deduplicate application events.
- First component: Deduplicate application events. Second component: Isolate tools behind a service layer.
- First component: Test interface-specific instruction behavior. Second component: Isolate tools behind a service layer.
- First component: Test interface-specific instruction behavior. Second component: Place durable behavior instructions in the system parameter.
Explanation
Correct Answer
C:The first proposal fits its assigned role because different Claude surfaces have different configuration and context mechanisms; an integration must supply its own instructions, while the second proposal fits its assigned role because a mediation service validates identity, arguments, authorization, and audit logs for each requested action; their respective acceptance criteria are to avoid assuming a Claude Code repository setting automatically affects a direct Messages API call and to avoid exposing raw production database credentials directly to an agent.
Incorrect Answers
A:The first proposal serves the different purpose to ensure an API assistant consistently uses a defined support role across requests instead of the assigned first role because system-level guidance supplies application-wide instruction context, separate from each user's task content, while the second proposal serves the different purpose to avoid starting two extraction jobs for the same uploaded document after repeated browser submissions instead of the assigned second role because an idempotent job key lets the server recognize duplicates at its request boundary; their respective acceptance criteria are to avoid assuming a Claude Code repository setting automatically affects a direct Messages API call and to avoid exposing raw production database credentials directly to an agent.
B:The first proposal serves the different purpose to avoid starting two extraction jobs for the same uploaded document after repeated browser submissions instead of the assigned first role because an idempotent job key lets the server recognize duplicates at its request boundary, while the second proposal fits its assigned role because a mediation service validates identity, arguments, authorization, and audit logs for each requested action; their respective acceptance criteria are to avoid assuming a Claude Code repository setting automatically affects a direct Messages API call and to avoid exposing raw production database credentials directly to an agent.
D:The first proposal fits its assigned role because different Claude surfaces have different configuration and context mechanisms; an integration must supply its own instructions, while the second proposal serves the different purpose to ensure an API assistant consistently uses a defined support role across requests instead of the assigned second role because system-level guidance supplies application-wide instruction context, separate from each user's task content; their respective acceptance criteria are to avoid assuming a Claude Code repository setting automatically affects a direct Messages API call and to avoid exposing raw production database credentials directly to an agent.
Question 51
In an agent platform project, acceptance test A checks whether a component can meet a requirement that proprietary repository contents remain within an approved execution environment. Acceptance test B checks whether a different component can build a coding agent that needs managed tool execution, sessions, and event handling. Which plan satisfies both tests?
- First component: Choose isolated sandbox execution for untrusted code. Second component: Use the Claude Agent SDK for an integrated tool-driven agent.
- First component: Deploy agent workers close to controlled data sources. Second component: Set a tool execution timeout.
- First component: Deploy agent workers close to controlled data sources. Second component: Use the Claude Agent SDK for an integrated tool-driven agent.
- First component: Set a tool execution timeout. Second component: Choose isolated sandbox execution for untrusted code.
Explanation
Correct Answer
C:The first proposal fits its assigned role because a self-hosted harness can enforce infrastructure, network, and retention controls for the surrounding tools, while the second proposal fits its assigned role because the SDK provides agent-loop primitives and integrations that remove much of the custom harness work; their respective acceptance criteria are to meet a requirement that proprietary repository contents remain within an approved execution environment and to build a coding agent that needs managed tool execution, sessions, and event handling.
Incorrect Answers
A:The first proposal serves the different purpose to let an agent run generated scripts without exposing production host files instead of the assigned first role because a restricted execution environment limits filesystem, network, and resource privileges, while the second proposal fits its assigned role because the SDK provides agent-loop primitives and integrations that remove much of the custom harness work; their respective acceptance criteria are to meet a requirement that proprietary repository contents remain within an approved execution environment and to build a coding agent that needs managed tool execution, sessions, and event handling.
B:The first proposal fits its assigned role because a self-hosted harness can enforce infrastructure, network, and retention controls for the surrounding tools, while the second proposal serves the different purpose to prevent a slow external inventory service from blocking an agent indefinitely instead of the assigned second role because deadlines bound latency and convert stalled calls into explicit recoverable failures; their respective acceptance criteria are to meet a requirement that proprietary repository contents remain within an approved execution environment and to build a coding agent that needs managed tool execution, sessions, and event handling.
D:The first proposal serves the different purpose to prevent a slow external inventory service from blocking an agent indefinitely instead of the assigned first role because deadlines bound latency and convert stalled calls into explicit recoverable failures, while the second proposal serves the different purpose to let an agent run generated scripts without exposing production host files instead of the assigned second role because a restricted execution environment limits filesystem, network, and resource privileges; their respective acceptance criteria are to meet a requirement that proprietary repository contents remain within an approved execution environment and to build a coding agent that needs managed tool execution, sessions, and event handling.
Question 52
An engineer extending an inference routing workflow has been asked to surface network timeouts differently from a completed response that declines a task. Which choice addresses the request most precisely?
- Separate transport failures from model refusals
- Use a queue for burst absorption
- Honor server rate-limit signals
- Parse responses by content block type
Explanation
Correct Answer
A:A transport error occurs before a valid application-level completion; refusals are model outcomes; this behavior directly supports the need to surface network timeouts differently from a completed response that declines a task. Treating both as one generic retry condition can loop endlessly.
Incorrect Answers
B:Queueing decouples producers from consumer throughput and enables controlled retries; this applies to the different goal to smooth sudden ingestion spikes before workers call a capacity-limited model API, not the requirement to surface network timeouts differently from a completed response that declines a task; increasing frontend timeouts does not eliminate backend capacity limits.
C:When an application needs to prevent a background processor from overwhelming the upstream API during bursts, throttling and backoff work with provider limits rather than assuming infinite request capacity; the current capability must instead surface network timeouts differently from a completed response that declines a task, while parallelism beyond quota can reduce successful throughput.
D:Structured outputs can contain different block kinds that need type-aware handling; the supported outcome is to avoid treating a tool call or thinking block as user-visible answer text, whereas this scenario requires the ability to surface network timeouts differently from a completed response that declines a task; concatenating arbitrary block payloads risks displaying internal control data.
Question 53
A design review identifies a missing capability in a conversation design project: it must show model-produced HTML snippets without allowing browser script injection. Which change meets that goal?
- Separate a partial streaming fragment from a finished object
- Escape generated content in the UI
- Parse JSON with a real parser
- Use bounded repair or regeneration
Explanation
Correct Answer
B:Output encoding or sanitization protects display contexts from untrusted markup; this behavior directly supports the need to show model-produced HTML snippets without allowing browser script injection. A correct JSON wrapper does not make embedded HTML safe.
Incorrect Answers
A:When an application needs to prevent the frontend from parsing an incomplete JSON chunk as a complete response, stream consumers must buffer and recognize completion before strict object validation; the current capability must instead show model-produced HTML snippets without allowing browser script injection, while an arbitrary content delta may end halfway through a field.
C:This approach serves the separate need to avoid accepting malformed generated objects because they resemble JSON to a reader because syntax validation converts serialized text into a trustworthy data structure only after parsing succeeds; it does not establish the outcome to show model-produced HTML snippets without allowing browser script injection, and regex extraction cannot generally validate nested JSON correctly.
D:A limited correction attempt can use validation errors as feedback before escalating; despite being useful when a service must recover malformed data without entering an infinite prompt-retry loop, it is not a solution to show model-produced HTML snippets without allowing browser script injection; unbounded retries can consume budget without resolving a defective input.
Question 54
A production API review asks the team to close two unrelated gaps without merging their execution paths. Service A must reject a numeric quote that is syntactically valid but outside the permitted range; service B must avoid burying the relevant contract clause beneath hundreds of low-value passages. Which paired remediation works?
- First component: Deduplicate application events. Second component: Use bounded retrieval results.
- First component: Validate semantic business constraints after parsing. Second component: Use bounded retrieval results.
- First component: Validate semantic business constraints after parsing. Second component: Keep user tasks in the user message.
- First component: Keep user tasks in the user message. Second component: Deduplicate application events.
Explanation
Correct Answer
B:The first proposal fits its assigned role because application logic must enforce business rules even after schema-valid output is produced, while the second proposal fits its assigned role because ranking and a context budget prioritize high-quality evidence while controlling input tokens; their respective acceptance criteria are to reject a numeric quote that is syntactically valid but outside the permitted range and to avoid burying the relevant contract clause beneath hundreds of low-value passages.
Incorrect Answers
A:The first proposal serves the different purpose to avoid starting two extraction jobs for the same uploaded document after repeated browser submissions instead of the assigned first role because an idempotent job key lets the server recognize duplicates at its request boundary, while the second proposal fits its assigned role because ranking and a context budget prioritize high-quality evidence while controlling input tokens; their respective acceptance criteria are to reject a numeric quote that is syntactically valid but outside the permitted range and to avoid burying the relevant contract clause beneath hundreds of low-value passages.
C:The first proposal fits its assigned role because application logic must enforce business rules even after schema-valid output is produced, while the second proposal serves the different purpose to separate an end user's request from the developer's stable operating instructions instead of the assigned second role because explicit message roles distinguish requested work from higher-priority operational framing; their respective acceptance criteria are to reject a numeric quote that is syntactically valid but outside the permitted range and to avoid burying the relevant contract clause beneath hundreds of low-value passages.
D:The first proposal serves the different purpose to separate an end user's request from the developer's stable operating instructions instead of the assigned first role because explicit message roles distinguish requested work from higher-priority operational framing, while the second proposal serves the different purpose to avoid starting two extraction jobs for the same uploaded document after repeated browser submissions instead of the assigned second role because an idempotent job key lets the server recognize duplicates at its request boundary; their respective acceptance criteria are to reject a numeric quote that is syntactically valid but outside the permitted range and to avoid burying the relevant contract clause beneath hundreds of low-value passages.
Question 55
A production-readiness check for a Claude application component requires the service to fail fast if a model ID or tool endpoint is absent in production. Which approach passes this check?
- Version prompt templates alongside code
- Validate configuration at startup
- Roll out configuration through reviewed changes
- Store API keys in managed secret storage
Explanation
Correct Answer
B:Startup checks surface missing required values before user traffic reaches a broken path; this behavior directly supports the need to fail fast if a model ID or tool endpoint is absent in production. A late runtime exception spreads one configuration defect across many requests.
Incorrect Answers
A:When an application needs to trace a quality regression to the exact instruction revision used, source-controlled prompt assets make rollbacks, diffs, and evaluations possible; the current capability must instead fail fast if a model ID or tool endpoint is absent in production, while editing prompts in an unversioned dashboard hides causal history.
C:This approach serves the separate need to prevent an untested tool-allowlist edit from reaching all developers immediately because change review and versioning provide accountability and a tested rollback path; it does not establish the outcome to fail fast if a model ID or tool endpoint is absent in production, and undocumented local overrides create configuration drift.
D:Central secret management supports scoped access, auditing, and key rotation; despite being useful when a service must rotate production credentials without committing new source code, it is not a solution to fail fast if a model ID or tool endpoint is absent in production; a plain-text config checked into a repository is an inappropriate credential store.
Question 56
An operations orchestration group is preparing a production change intended to answer routine account questions without the overhead of orchestrating multiple agents. Which technique matches that intent?
- Restrict each specialist's tool permissions
- Pass structured handoff artifacts between workers
- Keep a single agent with a small tool set
- Use an explicit completion condition and step budget
Explanation
Correct Answer
C:A single constrained executor simplifies state, authentication, logging, and latency where specialization adds little; this behavior directly supports the need to answer routine account questions without the overhead of orchestrating multiple agents. A multi-agent hierarchy is not a default requirement for tool use.
Incorrect Answers
A:This approach serves the separate need to allow a research worker to read records without granting its coordinator's write privileges because scoped tool capability prevents a delegated role from using authority unrelated to its assignment; it does not establish the outcome to answer routine account questions without the overhead of orchestrating multiple agents, and role descriptions alone do not revoke credentials available to the process.
B:Defined input and output schemas make it possible to attribute conclusions and detect missing fields at the handoff; despite being useful when a service must preserve evidence and data lineage when one agent delegates a subtask to another, it is not a solution to answer routine account questions without the overhead of orchestrating multiple agents; free-form chat summaries can omit identifiers required by the next worker.
D:A maximum iteration or cost budget plus a clear success criterion stops runaway tool loops; this applies to the different goal to bound the runtime of an agent that repeatedly searches without converging, not the requirement to answer routine account questions without the overhead of orchestrating multiple agents; merely asking the model to finish quickly does not enforce execution limits.
Question 57
A technical lead requests a solution that can separate an end user's request from the developer's stable operating instructions within a customer platform project. Which design should be chosen?
- Keep user tasks in the user message
- Separate model reasoning from authoritative calculations
- Create one session per user conversation
- Test interface-specific instruction behavior
Explanation
Correct Answer
A:Explicit message roles distinguish requested work from higher-priority operational framing; this behavior directly supports the need to separate an end user's request from the developer's stable operating instructions. Merging everything into one unstructured prompt weakens provenance.
Incorrect Answers
B:Deterministic services can enforce exact formulas and rounding before the model explains the result; the supported outcome is to compute discounts through verified business code rather than generated arithmetic alone, whereas this scenario requires the ability to separate an end user's request from the developer's stable operating instructions; a plausible model narrative is not a reliable financial ledger.
C:This approach serves the separate need to keep follow-up requests linked without exposing another tenant's messages because session scoping ensures history and tool results belong to the correct principal; it does not establish the outcome to separate an end user's request from the developer's stable operating instructions, and a global conversation list risks mixing private records.
D:Different Claude surfaces have different configuration and context mechanisms; an integration must supply its own instructions; despite being useful when a service must avoid assuming a Claude Code repository setting automatically affects a direct Messages API call, it is not a solution to separate an end user's request from the developer's stable operating instructions; files intended for a coding agent do not magically change API behavior.
Question 58
A customer platform team owns two independent components. The first must prevent bursts of document jobs from overwhelming rate limits and memory; the second must prevent a supposedly read-only status lookup from changing a server-side record. Which paired plan fulfills both?
- First component: Separate integration tests from unit tests. Second component: Avoid retrying non-idempotent writes blindly.
- First component: Avoid retrying non-idempotent writes blindly. Second component: Use explicit HTTP method semantics.
- First component: Enforce concurrency limits. Second component: Separate integration tests from unit tests.
- First component: Enforce concurrency limits. Second component: Use explicit HTTP method semantics.
Explanation
Correct Answer
D:The first proposal fits its assigned role because a semaphore or bounded queue keeps in-flight work within measured capacity, while the second proposal fits its assigned role because gET should retrieve resources while state changes use appropriate mutation methods and authorization checks; their respective acceptance criteria are to prevent bursts of document jobs from overwhelming rate limits and memory and to prevent a supposedly read-only status lookup from changing a server-side record.
Incorrect Answers
A:The first proposal serves the different purpose to verify that a real staging tool and its API contract work together instead of the assigned first role because integration tests exercise boundaries and serialization that isolated mocks may miss, while the second proposal serves the different purpose to keep an interrupted invoice-creation request from generating duplicate invoices instead of the assigned second role because retries of side-effecting actions need idempotency keys or transaction reconciliation; their respective acceptance criteria are to prevent bursts of document jobs from overwhelming rate limits and memory and to prevent a supposedly read-only status lookup from changing a server-side record.
B:The first proposal serves the different purpose to keep an interrupted invoice-creation request from generating duplicate invoices instead of the assigned first role because retries of side-effecting actions need idempotency keys or transaction reconciliation, while the second proposal fits its assigned role because gET should retrieve resources while state changes use appropriate mutation methods and authorization checks; their respective acceptance criteria are to prevent bursts of document jobs from overwhelming rate limits and memory and to prevent a supposedly read-only status lookup from changing a server-side record.
C:The first proposal fits its assigned role because a semaphore or bounded queue keeps in-flight work within measured capacity, while the second proposal serves the different purpose to verify that a real staging tool and its API contract work together instead of the assigned second role because integration tests exercise boundaries and serialization that isolated mocks may miss; their respective acceptance criteria are to prevent bursts of document jobs from overwhelming rate limits and memory and to prevent a supposedly read-only status lookup from changing a server-side record.
Question 59
The current design of a data pipeline pipeline does not adequately deploy a prompt change without exposing every customer to an untested regression. Which improvement resolves the gap?
- Implement automated regression tests
- Separate development, staging, and production environments
- Perform a threat review before go-live
- Define a staged release with rollback
Explanation
Correct Answer
D:Gradual rollout and versioned rollback allow metrics to be compared while limiting impact; this behavior directly supports the need to deploy a prompt change without exposing every customer to an untested regression. Editing the only production prompt in place makes reversions harder.
Incorrect Answers
A:When an application needs to catch schema and policy failures after a model or tool update, repeatable cases compare changes against documented acceptance criteria; the current capability must instead deploy a prompt change without exposing every customer to an untested regression, while a single successful demonstration provides little protection against regressions.
B:Environment boundaries keep credentials, data, and writes isolated during testing; the supported outcome is to exercise an agent's database tool without touching live customer records, whereas this scenario requires the ability to deploy a prompt change without exposing every customer to an untested regression; reusing production tokens in a local experiment defeats the separation.
C:This approach serves the separate need to uncover prompt-injection paths through retrieved webpages and connected tools because reviewing trust boundaries and permissions before deployment reveals attack routes unit tests may miss; it does not establish the outcome to deploy a prompt change without exposing every customer to an untested regression, and a post-incident review cannot prevent the initial exposure.
Question 60
An MCP interoperability solution has parallel requirements with separate owners. Workstream one must make one internal ticketing connector available to several agent clients; workstream two must give a coding agent a consistent review playbook and supporting assets. Which combination meets both?
- First component: Use hooks for enforceable safety checks. Second component: Package a Skill for reusable guidance.
- First component: Use a deterministic workflow for fixed stages. Second component: Use hooks for enforceable safety checks.
- First component: Publish an MCP server for reusable integrations. Second component: Use a deterministic workflow for fixed stages.
- First component: Publish an MCP server for reusable integrations. Second component: Package a Skill for reusable guidance.
Explanation
Correct Answer
D:The first proposal fits its assigned role because MCP standardizes tool and resource discovery across compatible applications, while the second proposal fits its assigned role because a Skill provides task procedures and resources that the agent can load when applicable; their respective acceptance criteria are to make one internal ticketing connector available to several agent clients and to give a coding agent a consistent review playbook and supporting assets.
Incorrect Answers
A:The first proposal serves the different purpose to deny destructive filesystem operations before a coding agent executes them instead of the assigned first role because runtime interception can apply deterministic policies to proposed actions, while the second proposal fits its assigned role because a Skill provides task procedures and resources that the agent can load when applicable; their respective acceptance criteria are to make one internal ticketing connector available to several agent clients and to give a coding agent a consistent review playbook and supporting assets.
B:The first proposal serves the different purpose to automate the same classify-review-export path across predictable submissions instead of the assigned first role because explicit application control makes each transition testable and easier to audit, while the second proposal serves the different purpose to deny destructive filesystem operations before a coding agent executes them instead of the assigned second role because runtime interception can apply deterministic policies to proposed actions; their respective acceptance criteria are to make one internal ticketing connector available to several agent clients and to give a coding agent a consistent review playbook and supporting assets.
C:The first proposal fits its assigned role because MCP standardizes tool and resource discovery across compatible applications, while the second proposal serves the different purpose to automate the same classify-review-export path across predictable submissions instead of the assigned second role because explicit application control makes each transition testable and easier to audit; their respective acceptance criteria are to make one internal ticketing connector available to several agent clients and to give a coding agent a consistent review playbook and supporting assets.