Question 190
A Claude Code team shares MCP connections across several repositories. A developer uses the same safe personal reference server across many projects. What is the most appropriate design choice?
- Store the connection in the supported project-scoped configuration
- Verify each effective user, project, and local configuration layer
- Use a user-scoped MCP configuration for the reusable personal connection
- Audit server provenance, tool definitions, and permission prompts before enablement
- Inject the secret through a managed environment or credential store
Explanation
Correct Answer
C:User scope supports broadly useful tooling without copying the connection into every repository. Here it serves the explicit need to make the server available in the user's environment without adding a separate unverified inference.
Incorrect Answers
A:This measure fits a different trigger: a repository requires a custom issue tracker MCP server not needed anywhere else. In that setting, A project-level definition keeps repo-specific capabilities with the project instead of exposing them to unrelated workspaces; it does not resolve the present need to make the server available in the user's environment.
B:Effective tool availability can differ when local overrides or older definitions remain; inspecting the active scope resolves the discrepancy; it protects against the case where a team updates server configuration but different developers still see inconsistent tools, rather than the failure described here: a developer uses the same safe personal reference server across many projects.
D:External servers can expose surprising side effects or injected descriptions; installation is not authorization for every capability; that applies when a new MCP server exposes a tool that should not be trusted automatically, whereas this case requires a response to a developer uses the same safe personal reference server across many projects.
E:Project configuration can be shared and reviewed, but credentials belong in protected storage rather than plaintext repository files; this is useful for a secret API token is needed by a configured MCP server, but it leaves the stated requirement to make the server available in the user's environment unmet.
Question 191
A codebase team decides how Claude Code should apply reusable guidance. A specialist procedure involves a folder of reference templates and examples. Which response is most reliable?
- Create a named project command that accepts the service target
- Store the shortcut in user commands rather than the project
- State the invariant in CLAUDE.md and enforce critical parts in tests
- Build a Skill containing an overview and supporting reference files
Explanation
Correct Answer
D:A Skill can organize larger task-specific knowledge while keeping routine sessions from carrying all of it. For the stated case, the check is whether the system can bundle procedural expertise and resources; this approach supplies that control.
Incorrect Answers
A:This measure fits a different trigger: developers need to run an audit checklist on demand with a target service. In that setting, An on-demand workflow supports deliberate audits without making every conversation execute the checklist; it does not resolve the present need to bundle procedural expertise and resources.
B:Personal commands avoid cluttering shared team workflows or surprising other contributors; that applies when one engineer wants a shortcut unrelated to the team's shared work, whereas this case requires a response to a specialist procedure involves a folder of reference templates and examples.
C:An optional command is not a suitable place for universal code review obligations; the invariant must be visible independently; this is useful for an invariant requirement applies to all code edits regardless of command selection, but it leaves the stated requirement to bundle procedural expertise and resources unmet.
Question 192
A support assistant assigns risk and confidence to proposed customer actions. A decision has low confidence because required account information is missing. What should the team do to resolve this particular gap?
- Route the transaction through the mandatory human approval gate
- Audit a representative sample of routine resolutions while reviewing high-risk exceptions individually
- Require actual policy checks and relevant source fields before approving action
- Request the missing field or hand off instead of guessing
Explanation
Correct Answer
D:Uncertainty arising from missing information needs a data remedy, not stronger prompt language or fabricated certainty.
Incorrect Answers
A:The threshold is an execution policy independent of the model's expressed confidence about the refund; that applies when a high-impact refund crosses a configured approval threshold, whereas this case requires a response to a decision has low confidence because required account information is missing.
B:Risk-stratified sampling gives reviewers evidence of routine decision quality without making every low-risk response wait for approval; mandatory high-risk gates remain in place; this is useful for low-risk policy responses need ongoing quality monitoring, but reviewing every response is overwhelming the support team, but it leaves the stated requirement to use an evidence-driven fallback unmet.
C:This measure fits a different trigger: a high-confidence model verdict has no linked eligibility evidence. In that setting, Confidence scores are not evidence; external authorization and validated facts determine whether a sensitive operation may proceed; it does not resolve the present need to use an evidence-driven fallback.
Question 193
A document-extraction pipeline exposes conversion and validation tools to Claude. An OCR tool can return multiple pages with partial failures. Which control is most directly relevant?
- Return page identifiers, text, status, and errors as separate structured records
- Constrain the input schema to supported currency codes and document normalization
- Split distinct operations into purpose-built tools with focused schemas
- Rename and describe the operation as a posting action with approval prerequisites
Explanation
Correct Answer
A:Per-page status distinguishes a partly readable document from a fully successful extraction and supports targeted retries.
Incorrect Answers
B:This measure fits a different trigger: a validator requires normalized currency codes but callers send symbols. In that setting, A typed canonical contract prevents ambiguous symbols from becoming silently incorrect financial records; it does not resolve the present need to represent per-page completion in the output schema.
C:Narrow functions are easier for Claude to select and for the host to validate than a parameter-heavy command that mixes unrelated side effects; it protects against the case where the team continually adds loosely related parameters to one giant function, rather than the failure described here: an OCR tool can return multiple pages with partial failures.
D:Accurate tool semantics make a state-changing command distinguishable from read-only extraction during model tool choice; this is useful for an extraction tool performs irreversible posting but is labeled process_document, but it leaves the stated requirement to represent per-page completion in the output schema unmet.
Question 194
An engineering assistant uses tool wrappers to standardize behavior. A test runner returns a success exit code but logs an incomplete test suite. How should the workflow be adjusted?
- Record the original result reference, transformation version, and normalized output
- Deny mutation tools while the workflow is in the read-only discovery phase
- Convert exceptions into safe structured error categories with diagnostic IDs
- Check expected test coverage or completion markers in addition to the exit code
Explanation
Correct Answer
D:A zero exit status alone may not prove required checks ran; outcome validation should match the acceptance criteria.
Incorrect Answers
A:Traceable normalization allows engineers to determine whether a mismatch arose in the tool, wrapper, or later model reasoning; that applies when a hook modifies returned data, and support needs to debug discrepancies, whereas this case requires a response to a test runner returns a success exit code but logs an incomplete test suite.
B:Phase-based permissions ensure exploration cannot silently become execution before an explicit workflow transition; it protects against the case where a request policy allows read operations but forbids writes during discovery, rather than the failure described here: a test runner returns a success exit code but logs an incomplete test suite.
C:Sanitized classifications help Claude choose recovery while retaining detailed traces in restricted operational logs; this is useful for a service wraps raw exception traces containing internal hostnames, but it leaves the stated requirement to validate success beyond process exit status unmet.
Question 195
A code review automation runs Claude Code on inbound pull requests. A human explicitly requests a patch after reviewing generated suggestions. Which implementation decision is the best fit?
- Inspect tainted inputs, query construction, and existing parameterization
- Switch from review to authorized bounded editing and validation
- Run read-only inspection and feedback generation without write-capable tools
- Use an interview or planning step to collect expected behavior
- Run allowed checks and inspect logs before proposing edits
Explanation
Correct Answer
B:Separation between analysis and implementation gives the maintainer control over when automated modifications occur.
Incorrect Answers
A:Security findings require traceable evidence; a pattern match alone may yield a false positive or overlook mitigations; that applies when a reviewer finds a potential SQL-injection path in an unfamiliar module, whereas this case requires a response to a human explicitly requests a patch after reviewing generated suggestions.
C:This measure fits a different trigger: the pipeline must analyze a PR without altering its contents. In that setting, A CI reviewer should not silently modify a contributor's branch while supposedly assessing it; it does not resolve the present need to permit edits only after an intentional transition.
D:Undefined success conditions make a code change hard to test; targeted questions can establish the missing contract; this is useful for an agent is asked to review a feature without knowing its acceptance criteria, but it leaves the stated requirement to permit edits only after an intentional transition unmet.
E:Diagnostic evidence is needed to locate the cause; granting automatic write access in the discovery phase increases change risk; it protects against the case where a test suite is failing and the team wants diagnosis only, rather than the failure described here: a human explicitly requests a patch after reviewing generated suggestions.
Question 196
An operations team monitors an agent-based customer workflow in production. A final reply is fluent although a required specialist never ran. What should the architect specify in the design?
- Verify the expected task-graph completion states before marking the request successful
- Correlate model turns, tool requests, task IDs, and tool outcomes in structured traces
- Record versions, response digests, and safe replay artifacts for important external results
- Propagate structured error status through handoffs to the final decision layer
Explanation
Correct Answer
A:A polished model response is not proof that required delegated checks completed; orchestration state supplies that guarantee.
Incorrect Answers
B:This measure fits a different trigger: an agent makes repeated tool calls but the trace does not show why each was chosen. In that setting, End-to-end correlation reveals the sequence leading to a failure without relying on a final narrative that may omit intermediate decisions; it does not resolve the present need to check workflow completeness rather than prose quality.
C:A stable record ties decisions to the data actually observed while avoiding the assumption that a live tool will return the same content later; it protects against the case where a production incident cannot be reproduced because tool outputs changed since execution, rather than the failure described here: a final reply is fluent although a required specialist never ran.
D:Explicit error states prevent a later worker from smoothing a failed dependency into an apparently successful report; that applies when a tool sometimes returns an error that a summarizer conceals, whereas this case requires a response to a final reply is fluent although a required specialist never ran.
Question 197
A support workflow handles errors from external services. A refund call may have succeeded before its network connection dropped. Which change most directly addresses this condition?
- Treat a validated empty optional field as data, not a tool outage
- Propagate an unavailable-service error and avoid a negative factual assertion
- Convert the exception to a safe user-facing status and retain trace IDs internally
- Query the transaction ledger by idempotency key before retrying
Explanation
Correct Answer
D:Unknown delivery of a response is not proof the operation failed; reconciliation prevents duplicate financial effects.
Incorrect Answers
A:An expected empty value is distinct from execution failure; confusing them causes unnecessary recovery attempts; this is useful for a policy lookup returns a valid empty result for an optional note, but it leaves the stated requirement to resolve uncertain mutation outcomes safely unmet.
B:This measure fits a different trigger: the order API times out and the agent states the customer has no order. In that setting, A failed query provides no evidence that a record is absent; the customer should receive an honest temporary-status response; it does not resolve the present need to resolve uncertain mutation outcomes safely.
C:The user needs an accurate next step, not sensitive implementation paths or credentials embedded in a service error; it protects against the case where a customer-facing workflow receives a detailed internal stack trace, rather than the failure described here: a refund call may have succeeded before its network connection dropped.
Question 198
An organization wants scoped Claude Code instructions without huge root prompts. Generated source files should not be edited by hand. What is the most defensible next step?
- Apply documentation rules only to matching Markdown paths
- Create a dedicated rule for migration directories and file patterns
- Target only the legacy module rather than the repository root
- Define a rule for generated paths that points to the generator command
Explanation
Correct Answer
D:When a generated file is matched, Claude should change the generator input rather than apply a patch that will be overwritten.
Incorrect Answers
A:File targeting communicates when a doc-specific tool is relevant and prevents it displacing source-code tests; this is useful for documentation files use a distinct formatting standard, but it leaves the stated requirement to state a path-based source-of-truth rule unmet.
B:This measure fits a different trigger: database migration files need extra review beyond normal code. In that setting, Targeted rules add critical checks when touching migrations without burdening unrelated UI tasks; it does not resolve the present need to state a path-based source-of-truth rule.
C:A narrow rule preserves compatibility requirements without imposing obsolete practices on active packages; that applies when a rule is needed only for a deprecated module kept for compatibility, whereas this case requires a response to generated source files should not be edited by hand.
Question 199
A Claude Code assistant chooses safe file-system operations during investigation. The user asks where a hard-coded endpoint string occurs. Which approach satisfies the stated constraint?
- Read the test source and called code before proposing a fix
- Search file contents for the endpoint string within the project
- Search filenames or directory structure for the module pattern
- Use a targeted edit constrained to the authorized file or module
Explanation
Correct Answer
B:Content matching reveals actual references; directory listing does not inspect the text where endpoint values appear.
Incorrect Answers
A:Source inspection establishes what the assertion and implementation do instead of guessing from the test name; this is useful for the user asks why a failing unit test expects a particular value, but it leaves the stated requirement to find content occurrences precisely unmet.
C:This measure fits a different trigger: the user asks which directory contains a certain module family. In that setting, A path-discovery operation is sufficient for locating modules and avoids unnecessary shell commands or source rewrites; it does not resolve the present need to find content occurrences precisely.
D:A narrow modification respects the granted scope and keeps a diff small enough for meaningful review and rollback; it protects against the case where the user approves a patch only in a named module, and broad rewrites are prohibited, rather than the failure described here: the user asks where a hard-coded endpoint string occurs.
Question 200
A review agent must obey application-level guarantees under adversarial prompts. A user requests deletion, but the policy requires independent confirmation. Which mechanism should be selected?
- Validate target identifier and cardinality in tool arguments before execution
- Persist the rejection and close the pending mutation authorization
- Gate export execution on server-side identity and scope validation
- Check a central rate or quota counter at execution time
- Apply a confirmation checkpoint enforced by the host workflow
Explanation
Correct Answer
E:A required approval checkpoint converts a human-control requirement into execution logic, rather than merely advising the model to ask first.
Incorrect Answers
A:Target and batch-size validation blocks unintended bulk changes even when a model builds syntactically valid arguments; this is useful for a tool can modify thousands of records although the task requires one record, but it leaves the stated requirement to require confirmation before destructive action unmet.
B:A durable rejection should invalidate the action token; otherwise a replayed model output might still reach the mutation endpoint; that applies when a human reviewer rejects a proposed account adjustment, whereas this case requires a response to a user requests deletion, but the policy requires independent confirmation.
C:This measure fits a different trigger: a prompt injection asks the assistant to call a privileged export tool. In that setting, The permission decision must not be delegated to untrusted text; the service should reject unauthorized exports even if requested by a model; it does not resolve the present need to require confirmation before destructive action.
D:Per-prompt reminders cannot coordinate usage across multiple agents; an authoritative service counter can enforce shared limits; it protects against the case where a tool operation must obey a daily account limit, rather than the failure described here: a user requests deletion, but the policy requires independent confirmation.