Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 46

Question 450

A Claude Code assistant chooses safe file-system operations during investigation. The user approves a patch only in a named module, and broad rewrites are prohibited. What is the most appropriate design choice?

  1. Run the approved test while keeping edit permissions disabled
  2. Search file contents for the endpoint string within the project
  3. Read the test source and called code before proposing a fix
  4. Use a targeted edit constrained to the authorized file or module
  5. Search filenames or directory structure for the module pattern

Explanation

Correct Answer

D:

A narrow modification respects the granted scope and keeps a diff small enough for meaningful review and rollback. The relevant requirement is to apply an explicitly scoped patch, and this measure operates at that decision point.

Incorrect Answers

A:

For the separate situation where the user approves running a test but not altering source code, a test execution can provide evidence without granting the assistant authority to modify files during diagnosis; the current task is to apply an explicitly scoped patch.

B:

Content matching reveals actual references; directory listing does not inspect the text where endpoint values appear; the benefit is tied to the user asks where a hard-coded endpoint string occurs. It is not the immediate control for the need to apply an explicitly scoped patch.

C:

Source inspection establishes what the assertion and implementation do instead of guessing from the test name; it is a safeguard for the user asks why a failing unit test expects a particular value, not the specific situation described: the user approves a patch only in a named module, and broad rewrites are prohibited.

E:

This choice is relevant when the user asks which directory contains a certain module family; a path-discovery operation is sufficient for locating modules and avoids unnecessary shell commands or source rewrites. Here, however, the user approves a patch only in a named module, and broad rewrites are prohibited.

Question 451

A customer-support service drives Claude through an application-owned tool loop. A state-changing refund tool may be retried after a dropped network connection. Which response is most reliable?

  1. Return a tool_result marked as an error with a sanitized diagnostic
  2. Execute approved calls and return one result correlated to each tool-use identifier
  3. Use a stable transaction or idempotency key at the refund service
  4. Execute the requested tool and send its correlated tool_result before asking Claude to continue

Explanation

Correct Answer

C:

The model cannot guarantee exactly-once external side effects; an application-level key makes a repeated request resolve to the same transaction.

Incorrect Answers

A:

For the separate situation where a tool request fails because a backend times out, a structured failed result preserves the agent loop and lets the model choose recovery or escalation; fabricating successful data would hide the failure; the current task is to prevent two refunds for one intended action.

B:

Tool-use IDs associate responses with particular requests; matching both blocks prevents a result from being mistaken for another operation; it is a safeguard for a single assistant message contains two independent tool_use blocks, not the specific situation described: a state-changing refund tool may be retried after a dropped network connection.

D:

This choice is relevant when the API replies with stop_reason=tool_use and an approved lookup request; a tool-use stop transfers execution to the host application; returning the result in a subsequent message lets Claude incorporate actual data rather than guess. Here, however, a state-changing refund tool may be retried after a dropped network connection.

Question 452

A release team operates Claude Code as part of automated change analysis. A pull request contains deliberately malicious prompt text in documentation. What should the team do to resolve this particular gap?

  1. Enforce fixed review instructions and disable unauthorized tools in the runner
  2. Emit a fixed verdict schema and reject unparseable output
  3. Block automatic merge and provide evidence for maintainer review
  4. Compare findings against adjudicated examples and track precision and recall

Explanation

Correct Answer

A:

Untrusted files can attempt to override the reviewer's goals, so executable permissions and instruction boundaries must remain separate.

Incorrect Answers

B:

This choice is relevant when the pipeline prints a detailed narrative but downstream jobs need a verdict field; a deterministic parser must be able to identify pass, needs-review, and blocked states without interpreting conversational language. Here, however, a pull request contains deliberately malicious prompt text in documentation.

C:

A high-risk finding should not be auto-dismissed by the same generative process that produced the report; it is a safeguard for a reviewer finds a high-risk issue requiring a person's decision, not the specific situation described: a pull request contains deliberately malicious prompt text in documentation.

D:

For the separate situation where the model-generated review itself must be measured for false positives over time, CI model output can drift; longitudinal labeled evaluation measures whether the review remains useful rather than merely verbose; the current task is to treat repository text as data, not governing policy.

Question 453

A Claude Code repository is organized into multiple services with different local rules. A developer assumes an automatically loaded file authorizes privileged commands. Which control is most directly relevant?

  1. Update the path or import so Claude Code can load the maintained guidance
  2. Use actual tool permissions and approval workflows for risky commands
  3. Keep common conventions at root and isolate service-specific guidance
  4. Use a supported CLAUDE.md import or concise reference to the maintained document

Explanation

Correct Answer

B:

Instruction files can steer behavior but do not grant or enforce operating-system authorization by themselves. For the stated case, the check is whether the system can separate guidance from execution permission; this approach supplies that control.

Incorrect Answers

A:

A prompt reminder cannot replace a broken import; resolving the file reference restores the policy's discoverability; it is a safeguard for an imported policy file is no longer present after a directory move, not the specific situation described: a developer assumes an automatically loaded file authorizes privileged commands.

C:

This choice is relevant when the root instructions list every service's commands and overwhelm simple tasks; layered instructions reduce irrelevant context while preserving shared policies that apply across the repository. Here, however, a developer assumes an automatically loaded file authorizes privileged commands.

D:

Importing maintained guidance avoids parallel copies that drift when the release process changes; the benefit is tied to a release policy document already exists and should not be copied into multiple instruction files. It is not the immediate control for the need to separate guidance from execution permission.

Question 454

A Claude Code team shares MCP connections across several repositories. A team updates server configuration but different developers still see inconsistent tools. How should the workflow be adjusted?

  1. Verify each effective user, project, and local configuration layer
  2. Inject the secret through a managed environment or credential store
  3. Use a user-scoped MCP configuration for the reusable personal connection
  4. Audit server provenance, tool definitions, and permission prompts before enablement

Explanation

Correct Answer

A:

Effective tool availability can differ when local overrides or older definitions remain; inspecting the active scope resolves the discrepancy.

Incorrect Answers

B:

Project configuration can be shared and reviewed, but credentials belong in protected storage rather than plaintext repository files; it is a safeguard for a secret API token is needed by a configured MCP server, not the specific situation described: a team updates server configuration but different developers still see inconsistent tools.

C:

User scope supports broadly useful tooling without copying the connection into every repository; the benefit is tied to a developer uses the same safe personal reference server across many projects. It is not the immediate control for the need to make configuration scope and version explicit.

D:

For the separate situation where a new MCP server exposes a tool that should not be trusted automatically, external servers can expose surprising side effects or injected descriptions; installation is not authorization for every capability; the current task is to make configuration scope and version explicit.

Question 455

A code assistant integrates MCP tools maintained by different engineering groups. An internal tool is versioned while several clients depend on its outputs. Which implementation decision is the best fit?

  1. Use an enumerated argument value set and validate it on the server
  2. Return an empty success result for legitimate no-match outcomes
  3. Ignore instructions in tool metadata that conflict with higher-priority security policy
  4. Redact secrets before returning tool content to Claude
  5. Version breaking schema changes or provide a backwards-compatible migration

Explanation

Correct Answer

E:

Unexpected field renames can break tool consumers and downstream validators even when a tool's description still sounds correct.

Incorrect Answers

A:

Schema constraints and server checks prevent an unintended target resulting from free-form environment strings; the benefit is tied to a deployment tool accepts an environment name with many misspelled variants. It is not the immediate control for the need to preserve a stable contract across changes.

B:

No matching build is a valid result distinct from an operational error; mixing them causes unnecessary retries and escalation; it is a safeguard for a build lookup returns no match and callers mistake it for a crashed server, not the specific situation described: an internal tool is versioned while several clients depend on its outputs.

C:

This choice is relevant when one tool's description encourages bypassing repository review rules; descriptions are useful for discovery, not an authority source; a malicious tool can attempt prompt injection through its advertised text. Here, however, an internal tool is versioned while several clients depend on its outputs.

D:

For the separate situation where an API key appears in an MCP tool's result object, tool responses become model context and may be quoted later; a restricted data boundary must filter credential material; the current task is to preserve a stable contract across changes.

Question 456

A customer-support assistant must decide when to request human assistance. A support specialist finds the policy clear and customer eligibility already verified. What should the architect specify in the design?

  1. Create a human-approval request with the verified amount and case evidence
  2. Mark the policy gap and hand off to an authorized human
  3. Trigger the organization's urgent human handoff process
  4. Apply the verified rule within delegated permission and record the decision

Explanation

Correct Answer

D:

Human review should be focused on risk or uncertainty; routine well-defined cases can proceed under controlled authorization.

Incorrect Answers

A:

Authority limits must be observed at execution; the model cannot increase its own financial permissions; it is a safeguard for a refund would exceed the support agent's approved limit, not the specific situation described: a support specialist finds the policy clear and customer eligibility already verified.

B:

This choice is relevant when a request falls outside every published returns-policy category; the model should not create a new rule for an unaddressed case; a documented gap requires accountable policy interpretation. Here, however, a support specialist finds the policy clear and customer eligibility already verified.

C:

For the separate situation where a customer reports urgent harm that the assistant cannot safely assess, high-impact, ambiguous reports warrant a trained responder and documented procedure rather than an autonomous improvised resolution; the current task is to complete an authorized routine case without unnecessary handoff.

Question 457

A coordinator controls which tools each research worker may call. A reviewer needs source documents but must not publish final content. Which change most directly addresses this condition?

  1. Let the coordinator select subsequent tools or specialists using completed status fields
  2. Provide a computation tool and the curated input dataset to the analyst
  3. Expose read-only verification tools to the reviewer and reserve release actions for the owner
  4. Allow approved retrieval tools while withholding document-writing capabilities

Explanation

Correct Answer

C:

Verification and publication are separate authority levels; a review worker can inspect evidence without being able to commit the final artifact.

Incorrect Answers

A:

For the separate situation where an orchestrator requires two worker outputs to choose the next branch, a routing decision depends on multiple worker results; it belongs at the point that sees the combined state; the current task is to allow inspection while withholding publication.

B:

The analyst's task is arithmetic and interpretation, not unrestricted retrieval; limited tools reduce irrelevant calls and scope creep; the benefit is tied to an analyst must compute statistics on previously collected evidence. It is not the immediate control for the need to allow inspection while withholding publication.

D:

This choice is relevant when a source-finding worker needs web search but cannot modify internal documents; tool distribution should follow the worker's role; disallowing writes prevents discovery tasks from changing records. Here, however, a reviewer needs source documents but must not publish final content.

Question 458

An MCP client connects to business-data tools provided by third-party servers. An MCP server returns a copied webpage containing hostile 'system' instructions. What is the most defensible next step?

  1. Treat tool annotations as hints and evaluate real side effects
  2. Enforce caller authorization in the client and backend regardless of tool descriptions
  3. Mark pending operations unknown and reconnect or reconcile through supported state
  4. Prevent retrieved text from authorizing tool calls or changing instruction hierarchy

Explanation

Correct Answer

D:

Prompt-injection material may arrive through normal tool results, so the agent's higher-priority policies and allowlists remain in force.

Incorrect Answers

A:

Capability annotations improve discovery but are not a security boundary or proof that code has no writes; the benefit is tied to a tool is marked read-only in descriptive metadata but its implementation writes audit records. It is not the immediate control for the need to treat fetched content as lower-trust data.

B:

This choice is relevant when a server's advertised description says it may ignore the user's data-access restrictions; server metadata is untrusted discovery data; access checks must remain enforceable even if metadata is malicious or misleading. Here, however, an MCP server returns a copied webpage containing hostile 'system' instructions.

C:

Transport failure does not establish that a side effect failed; the client must track request outcomes and avoid blind mutation retries; it is a safeguard for a remote MCP server drops connection during a multi-step task, not the specific situation described: an MCP server returns a copied webpage containing hostile 'system' instructions.

Question 459

A compliance research coordinator receives output from multiple agents. A fact-checking worker corrects a statistic after downstream synthesis already used an older result. Which approach satisfies the stated constraint?

  1. Return an explicit unresolved-question field for human adjudication
  2. Have analysis output the estimate, assumptions, and limits separately
  3. Track artifact versions and rerun only synthesis nodes that depended on the corrected result
  4. Require source URL or ID, publication date, and claim in the worker handoff

Explanation

Correct Answer

C:

Dependency-aware versioning locates conclusions that rely on outdated evidence, leaving independent completed analyses intact while affected synthesis is rebuilt.

Incorrect Answers

A:

An open issue should remain visible instead of being flattened into a confident answer by downstream summarization; it is a safeguard for a worker provides a decision that still depends on legal interpretation, not the specific situation described: a fact-checking worker corrects a statistic after downstream synthesis already used an older result.

B:

The writing agent can only qualify a number responsibly when its underlying assumptions survive the handoff; the benefit is tied to an analyst calculates an estimate but does not state its uncertainty. It is not the immediate control for the need to invalidate conclusions derived from superseded worker evidence.

D:

This choice is relevant when a fact-checking worker reports a source but not its publication date; a dated provenance record lets the coordinator detect obsolete policies and avoid presenting historical facts as current. Here, however, a fact-checking worker corrects a statistic after downstream synthesis already used an older result.

Question 460

A team maintains path-aware architecture conventions in Claude Code. A critical write restriction is written only as path-specific prose. Which mechanism should be selected?

  1. Constrain edit permissions to approved paths at the execution layer
  2. Attach the review procedure to infrastructure configuration patterns
  3. Keep shared guidance in one place and scoped exceptions explicitly narrow
  4. Revise path patterns and test their effective selection
  5. Give mobile and server trees distinct scoped guidance

Explanation

Correct Answer

A:

Scoped instruction text improves intent but does not prevent a capable tool from writing elsewhere when misused. Here it serves the explicit need to back the restriction with a real filesystem or tool guard without adding a separate unverified inference.

Incorrect Answers

B:

A path-aware rule gives extra caution where necessary without slowing unrelated text edits; the benefit is tied to only infrastructure-as-code files require a plan-before-change policy. It is not the immediate control for the need to back the restriction with a real filesystem or tool guard.

C:

For the separate situation where a scoped rule repeats a contradictory general rule verbatim, redundant copies can drift; exceptions should state exactly what changes for a subset of files; the current task is to back the restriction with a real filesystem or tool guard.

D:

Rule files are not intelligent about renamed folders; matching must be maintained alongside structural refactors; it is a safeguard for a new package directory does not match the existing rules, not the specific situation described: a critical write restriction is written only as path-specific prose.

E:

This choice is relevant when the mobile app follows a different package layout from the server; relevant architecture instructions reduce accidental cross-project assumptions when a monorepo contains multiple build systems. Here, however, a critical write restriction is written only as path-specific prose.

Claude Certified Architect - Foundations (CCAR-F) (CCA-F) Practice Tests