Question 20
A regulated advice assistant lacks enough verified evidence to make a material recommendation. What should the team do at this decision point?
- Provide an escalation channel for uncertain outcomes
- Show progress for asynchronous jobs
- Clearly distinguish a draft from a committed action
- Use CLAUDE.md for repository guidance in Claude Code
Explanation
Correct Answer
A:Escalation transfers context to an authorized person instead of fabricating a final decision, although human review must be a real operational path, not only a sentence in a prompt.
Incorrect Answers
B:Progress state and job identifiers give a reliable view of work in flight; however, the critical mechanism here is that escalation transfers context to an authorized person instead of fabricating a final decision, and a fake static percentage would misrepresent actual completion.
C:Labels and separate commit controls prevent users from assuming a draft has been submitted; the requirement instead turns on this behavior: escalation transfers context to an authorized person instead of fabricating a final decision; model-generated wording should not make a simulated operation appear executed.
D:Repository instruction files provide persistent project context to the Claude Code environment; this cannot substitute for the behavior that escalation transfers context to an authorized person instead of fabricating a final decision, because they are not automatically loaded into unrelated Messages API clients.
Question 21
A model produces a plausible shell command during a troubleshooting chat and the product must avoid executing it blindly. What change is best aligned with the goal?
- Treat retrieved pages as untrusted evidence
- Separate display text from operational commands
- Separate code execution from generated text
- Keep user instructions subordinate to security policy
Explanation
Correct Answer
C:Text suggestions remain inert until authorized application logic validates and runs an approved operation; a crucial qualification is that rendering a command in the answer is not an execution authorization.
Incorrect Answers
A:External page text should be handled as data, not as new operational authority; this addresses a separate concern, not the fact that text suggestions remain inert until authorized application logic validates and runs an approved operation; a page author cannot legitimately override the application's system policy.
B:The service distinguishes a human-readable recommendation from an authorized tool action; here, the controlling capability is that text suggestions remain inert until authorized application logic validates and runs an approved operation, but parsing ordinary prose to infer permission for a write is unsafe.
D:The application enforces identity and authorization checks despite the request's wording; by comparison, the intended solution relies on this: text suggestions remain inert until authorized application logic validates and runs an approved operation; requests from end users do not grant new backend privileges.
Question 22
Repository maintainers want a consistent release checklist that a coding agent can apply on demand. What is the appropriate next design choice?
- Use a subagent for context-isolated auditing
- Create a project Skill for recurring checks
- Use a hook for mandatory pre-execution validation
- Use a Skill for reusable procedures
Explanation
Correct Answer
B:Skill content packages reusable instructions and supporting guidance for the task; the important constraint is that the project must still provide actual tests and tools needed to perform the checks.
Incorrect Answers
A:A delegated worker returns focused findings while keeping its verbose exploration isolated, whereas the requirement calls for another behavior: skill content packages reusable instructions and supporting guidance for the task; separate context is not a substitute for least-privilege tool access.
C:A lifecycle hook evaluates the proposed action outside the model and can block it deterministically; the requirement instead turns on this behavior: skill content packages reusable instructions and supporting guidance for the task; natural-language checklists are weaker than enforced execution controls.
D:Skills package instructional workflows that an agent can reuse within supported clients; this cannot substitute for the behavior that skill content packages reusable instructions and supporting guidance for the task, because they are guidance, not network APIs or execution permissions.
Question 23
The organization observes inference calls at unusual hours and wants to catch credential misuse early. How should the implementation handle this need?
- Check end-user permissions before using a customer tool
- Rotate a key after suspected leakage
- Store credentials in a secret manager
- Monitor key usage anomalies
Explanation
Correct Answer
D:Usage attribution, budget alerts, and access logs can expose unauthorized patterns for investigation; its operational boundary is that a key that still works is not evidence that its use is legitimate.
Incorrect Answers
A:The backend verifies the caller is allowed to access that specific object; however, the critical mechanism here is that usage attribution, budget alerts, and access logs can expose unauthorized patterns for investigation, and a valid session establishes identity but not ownership of every possible resource.
B:Revoking and replacing the affected secret reduces the time attackers can continue using it; the requirement instead turns on this behavior: usage attribution, budget alerts, and access logs can expose unauthorized patterns for investigation; hiding the log page alone cannot guarantee the secret was not captured.
C:Protected storage and runtime injection keep billable API keys out of committed files; this cannot substitute for the behavior that usage attribution, budget alerts, and access logs can expose unauthorized patterns for investigation, because encryption at rest does not justify exposing secrets in source control.
Question 24
A client sends tool arguments that may contain quotes, line breaks, and Unicode text. What is the right implementation response?
- Use typed response variants where applicable
- Escape output for its destination context
- Deserialize JSON with expected schema checks
- Encode JSON using a proper serializer
Explanation
Correct Answer
D:A JSON encoder produces correctly escaped structured payloads; the implementation must still recognize that manually concatenated JSON strings are vulnerable to malformed syntax and injection bugs.
Incorrect Answers
A:A discriminated union or type dispatch handles text, tool calls, and other supported blocks correctly; this cannot substitute for the behavior that a JSON encoder produces correctly escaped structured payloads, because treating all blocks as plain strings discards structural information.
B:Context-aware escaping prevents text from becoming active script or HTML; in contrast, a JSON encoder produces correctly escaped structured payloads, while JSON escaping is not interchangeable with HTML or SQL escaping.
C:Typed validation enforces the documented contract after parsing; this addresses a separate concern, not the fact that a JSON encoder produces correctly escaped structured payloads; JSON syntax validity does not prove field meaning or user authorization.
Question 25
A system must reason through difficult multi-file code migrations and complex failure interactions. Which change would satisfy the constraint?
- Prefer Sonnet-class models for balanced workloads
- Route dynamically among evaluated tiers
- Use a fast tier for request classification
- Prefer Opus-class models for demanding reasoning
Explanation
Correct Answer
D:The top capability tier is a candidate when advanced reasoning quality dominates the tradeoff—with this limitation: higher cost and time should be justified through measured improvement.
Incorrect Answers
A:A balanced model tier can provide strong capability at lower overhead than the highest tier; the needed capability is different—the top capability tier is a candidate when advanced reasoning quality dominates the tradeoff—and higher complexity may still justify a more powerful model.
B:A routing policy can send easy cases to economical models and escalate complex work, whereas the requirement calls for another behavior: the top capability tier is a candidate when advanced reasoning quality dominates the tradeoff; misclassification by the router can lower quality on unexpectedly difficult tasks.
C:A small capable classifier avoids spending top-tier reasoning cost on simple routing; however, the critical mechanism here is that the top capability tier is a candidate when advanced reasoning quality dominates the tradeoff, and poor routing accuracy can still misdirect expensive follow-up steps.
Question 26
A caller reaches a private-account assistant without a verified session, and the service must establish who is acting before evaluating account-level permissions. What is the best fit for the described situation?
- Authorize access to stored conversation history
- Authorize each requested operation at the resource level
- Audit privileged changes with actor attribution
- Authenticate the caller before processing account actions
Explanation
Correct Answer
D:Identity verification establishes who is requesting the action before tool access is considered; a convincing statement of ownership inside a prompt is not authentication.
Incorrect Answers
A:Access-control checks at the storage layer prevent cross-account conversation disclosure; here, the controlling capability is that identity verification establishes who is requesting the action before tool access is considered, but a hidden session identifier is not enough to authorize retrieval.
B:Access policies check whether that principal may perform the exact requested change on the target object; the missing capability is more specific: identity verification establishes who is requesting the action before tool access is considered; successful login alone does not grant permission across all records.
C:An audit trail connects the principal, approved operation, target, and result; by comparison, the intended solution relies on this: identity verification establishes who is requesting the action before tool access is considered; a text summary by the model is not sufficient evidence of a committed security change.
Question 27
A workflow succeeds on one workstation but fails on fresh machines due to an undocumented runtime dependency. Which solution matches the actual constraint?
- Store environment-specific endpoints in configuration
- Check effective settings across scopes
- Document required environment variables
- Fail closed on missing safety configuration
Explanation
Correct Answer
C:An explicit configuration contract makes setup repeatable without embedding secret values; notably, copying the original developer's entire environment is neither secure nor maintainable.
Incorrect Answers
A:Deployment configuration injects environment values without rewriting application logic; however, the critical mechanism here is that an explicit configuration contract makes setup repeatable without embedding secret values, and hardcoded production URLs increase the chance of accidental test writes.
B:Resolving effective configuration reveals higher-priority policy and machine-specific overrides; the requirement instead turns on this behavior: an explicit configuration contract makes setup repeatable without embedding secret values; reading only one settings file can miss the operative restriction.
D:Startup validation blocks unsafe execution rather than falling back to all permissions; in contrast, an explicit configuration contract makes setup repeatable without embedding secret values, while the absence of a guardrail must not silently widen access.
Question 28
An assistant needs to query legacy ticket records with an unfamiliar internal protocol. Which approach best meets this requirement?
- Use a dedicated tool adapter for private systems
- Use a background job worker for long runs
- Use an event-driven trigger for external updates
- Validate generated tool parameters
Explanation
Correct Answer
A:An adapter exposes narrowly typed operations and translates requests into the legacy interface; the important constraint is that the model does not acquire direct private-network access by seeing the system's name.
Incorrect Answers
B:Worker queues provide durable processing and status reporting outside the initial HTTP connection; the requirement instead turns on this behavior: an adapter exposes narrowly typed operations and translates requests into the legacy interface; synchronous endpoints can time out under long processing loads.
C:Event handlers initiate the approved model workflow on a real business change; this cannot substitute for the behavior that an adapter exposes narrowly typed operations and translates requests into the legacy interface, because polling or repeated user prompts are unnecessary when a reliable event is available.
D:Server-side validation constrains types, ranges, ownership, and supported operations; in contrast, an adapter exposes narrowly typed operations and translates requests into the legacy interface, while a well-described function schema does not replace authorization checks.
Question 29
An assistant receives an authenticated request to open a file that belongs to another account. Which capability is necessary for this outcome?
- Monitor key usage anomalies
- Check end-user permissions before using a customer tool
- Audit approval decisions for high-impact actions
- Revoke integration access when offboarding
Explanation
Correct Answer
B:The backend verifies the caller is allowed to access that specific object; in production, a valid session establishes identity but not ownership of every possible resource.
Incorrect Answers
A:Usage attribution, budget alerts, and access logs can expose unauthorized patterns for investigation; the missing capability is more specific: the backend verifies the caller is allowed to access that specific object; a key that still works is not evidence that its use is legitimate.
C:A durable record captures the approver, scope, timestamp, and action approved; the needed capability is different—the backend verifies the caller is allowed to access that specific object—and an assistant's paraphrase of approval is not equivalent to an auditable authorization event.
D:Offboarding revokes credentials and grants to prevent future unauthorized use, whereas the requirement calls for another behavior: the backend verifies the caller is allowed to access that specific object; deleting the employee's saved chat does not revoke external tool credentials.
Question 30
A remote client asks to modify a record it does not own but supplies syntactically valid arguments. Which technical control is appropriate?
- Version and test server interfaces
- Use stdio for a local subprocess integration
- Validate operations in the server handler
- Advertise tool metadata clearly
Explanation
Correct Answer
C:Server handlers enforce identity, scope, and business rules before accessing protected systems, subject to this requirement: trusting the caller's description is not a security control.
Incorrect Answers
A:Compatibility tests and versioned rollout protect consumers from unexpected protocol or field changes; this cannot substitute for the behavior that server handlers enforce identity, scope, and business rules before accessing protected systems, because unreviewed changes can break integrations even when the server starts successfully.
B:Standard input and output suit tightly coupled local client-server interactions; in contrast, server handlers enforce identity, scope, and business rules before accessing protected systems, while the server's standard output must remain protocol-clean instead of mixing human debug prints.
D:The server supplies discoverable names, descriptions, and input schemas for the tool set; here, the controlling capability is that server handlers enforce identity, scope, and business rules before accessing protected systems, but misleading descriptions can cause inappropriate model tool selection.