Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 23

Question 220

An MCP service encounters errors while fetching customer records. A database loses connectivity during a lookup. How should the workflow be adjusted?

  1. Return a permission-denied outcome and require an authorized path
  2. Return a structured validation error identifying the affected field
  3. Classify the issue as transient unavailable and provide a safe diagnostic identifier
  4. Return a distinct not-found status without recommending blind retries
  5. Return a structured rate-limit error and retry-after guidance

Explanation

Correct Answer

C:

Clear error categorization lets orchestration retry selectively while keeping internal stack traces out of model-facing content.

Incorrect Answers

A:

This measure fits a different trigger: a tool call is rejected because the caller lacks access to the account. In that setting, A permission error cannot be fixed by repeating the same call; it needs identity correction or escalation without leaking account details; it does not resolve the present need to identify a recoverable service failure.

B:

When an invalid date format fails argument validation before tool execution, this would be useful because a field-specific error supports a corrected retry, whereas generic failure text makes it hard to choose an appropriate next action; this item instead calls for a way to identify a recoverable service failure.

D:

A genuinely missing resource is different from transient infrastructure failure; retrying unchanged identifiers adds load without resolving it; the alternative would make sense if a requested account number is well formed but does not exist. It does not directly fix the observed condition: a database loses connectivity during a lookup.

E:

Throttling is transient and should be handled with bounded delay, not treated as an absent customer record; the relevant precondition would be a backend returns HTTP 429 with a retry window. The question instead describes a database loses connectivity during a lookup.

Question 221

A Claude Code assistant chooses safe file-system operations during investigation. The user asks why a failing unit test expects a particular value. Which implementation decision is the best fit?

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

Explanation

Correct Answer

C:

Source inspection establishes what the assertion and implementation do instead of guessing from the test name. This matches the workflow goal to inspect the specific test and nearby implementation rather than introducing an unrelated action.

Incorrect Answers

A:

Content matching reveals actual references; directory listing does not inspect the text where endpoint values appear; the alternative would make sense if the user asks where a hard-coded endpoint string occurs. It does not directly fix the observed condition: the user asks why a failing unit test expects a particular value.

B:

This measure fits a different trigger: the user approves running a test but not altering source code. In that setting, A test execution can provide evidence without granting the assistant authority to modify files during diagnosis; it does not resolve the present need to inspect the specific test and nearby implementation.

D:

When the user approves a patch only in a named module, and broad rewrites are prohibited, this would be useful because a narrow modification respects the granted scope and keeps a diff small enough for meaningful review and rollback; this item instead calls for a way to inspect the specific test and nearby implementation.

Question 222

An MCP client connects to business-data tools provided by third-party servers. A remote MCP server drops connection during a multi-step task. What should the architect specify in the design?

  1. Prevent retrieved text from authorizing tool calls or changing instruction hierarchy
  2. Treat tool annotations as hints and evaluate real side effects
  3. Bind identity and allowed resources at the server execution boundary
  4. Mark pending operations unknown and reconnect or reconcile through supported state

Explanation

Correct Answer

D:

Transport failure does not establish that a side effect failed; the client must track request outcomes and avoid blind mutation retries.

Incorrect Answers

A:

When an MCP server returns a copied webpage containing hostile 'system' instructions, this would be useful because prompt-injection material may arrive through normal tool results, so the agent's higher-priority policies and allowlists remain in force; this item instead calls for a way to handle transport interruption without inventing results.

B:

Capability annotations improve discovery but are not a security boundary or proof that code has no writes; the alternative would make sense if a tool is marked read-only in descriptive metadata but its implementation writes audit records. It does not directly fix the observed condition: a remote MCP server drops connection during a multi-step task.

C:

This measure fits a different trigger: multiple users connect to the same server with different data permissions. In that setting, A shared server must not conflate clients; prompt context and tool-list filtering alone are insufficient to isolate data access; it does not resolve the present need to handle transport interruption without inventing results.

Question 223

An agentic research pipeline must propagate failures through several workers. A writing agent receives an upstream timeout labeled success. Which change most directly addresses this condition?

  1. Pass a typed timeout outcome instead of a successful empty payload
  2. Mark the calculation blocked and include the validation failure
  3. Return per-source status with valid outputs and a partial-completion flag
  4. Refuse publication until mandatory worker outcomes are successful or formally waived

Explanation

Correct Answer

A:

Correct error semantics prevent downstream agents from mistaking unavailable data for evidence of absence. This matches the workflow goal to fix error propagation at the source boundary rather than introducing an unrelated action.

Incorrect Answers

B:

A computed value obtained from unverified inputs should not be promoted as evidence merely because a number was produced; the alternative would make sense if an analysis agent reports a numeric result but its input validation failed. It does not directly fix the observed condition: a writing agent receives an upstream timeout labeled success.

C:

This measure fits a different trigger: one of three independent data sources fails while the others are usable. In that setting, An all-or-nothing summary discards useful work, while an all-success label conceals missing coverage; partial status preserves both truths; it does not resolve the present need to fix error propagation at the source boundary.

D:

When the final publication stage needs all required source checks complete, this would be useful because a release gate should inspect explicit prerequisite state rather than trust an optimistic synthesis paragraph; this item instead calls for a way to fix error propagation at the source boundary.

Question 224

A Claude Code automation project uses lifecycle hooks for deterministic behavior. A session needs to load its approved repository-specific working conventions. What is the most defensible next step?

  1. Sanitize tool outputs at a controlled result boundary before passing them to the model
  2. Capture the subagent lifecycle completion event for orchestration tracking
  3. Use session initialization to provide relevant instructions or environment facts
  4. Run the targeted test through an appropriate post-tool hook or workflow stage

Explanation

Correct Answer

C:

Start-of-session hooks can make standard setup repeatable rather than relying on each developer to manually paste conventions.

Incorrect Answers

A:

When a tool result contains a bearer token accidentally echoed from logs, this would be useful because tool-return filtering reduces exposure of secrets in model context and later generated text; a user-facing disclaimer does not remove the token; this item instead calls for a way to initialize known context at session start.

B:

This measure fits a different trigger: a subagent finishes a task, and its completion status must be recorded centrally. In that setting, A completion event is the right place to record the specialist's result without confusing it with the parent agent's final turn; it does not resolve the present need to initialize known context at session start.

D:

The validation is bound to the actual modification event, not to whether Claude remembers a reminder at the end of the conversation; the relevant precondition would be a test command must run after a source-editing tool successfully changes code. The question instead describes a session needs to load its approved repository-specific working conventions.

Question 225

An engineering company integrates Claude Code into a secure CI workflow. A nightly review compares generated findings with a baseline. Which approach satisfies the stated constraint?

  1. Use separate restricted runners or permission scopes for untrusted PR analysis
  2. Gate success on actual test and lint exit statuses, not the assistant's assertion
  3. Grant only the API permission required for review comments and withhold repository-write capability
  4. Record the revision, prompt configuration, and tool versions for each run
  5. Use schema-checked JSON findings with severity and source locations

Explanation

Correct Answer

D:

Traceable inputs help explain why findings changed and distinguish code changes from changes in review configuration.

Incorrect Answers

A:

When the reviewer has access to production deployment credentials through the runner image, this would be useful because a model that reads untrusted code should not share a security boundary with credentials capable of changing production; this item instead calls for a way to retain reproducible review inputs.

B:

This measure fits a different trigger: a build reports a Claude-generated pass despite skipped unit tests. In that setting, A generated summary is evidence to inspect, not a replacement for deterministic pipeline results; it does not resolve the present need to retain reproducible review inputs.

C:

Least privilege lets CI publish feedback without creating commits or changing protected source files; the relevant precondition would be the job must comment on a PR but may not modify the branch. The question instead describes a nightly review compares generated findings with a baseline.

E:

A validated contract lets downstream code render and deduplicate findings without trusting prose extraction; the alternative would make sense if the pipeline requires a stable format for severity, path, and line number. It does not directly fix the observed condition: a nightly review compares generated findings with a baseline.

Question 226

A document-processing architecture decides between online calls and batch jobs. A batch worker may retry failed items after network interruption. Which mechanism should be selected?

  1. Use batch requests with durable tracking and completion monitoring
  2. Invoke the model synchronously with bounded timeout and fallback
  3. Compute field-level error metrics by input cohort and workflow version
  4. Maintain stable item IDs and a durable submission ledger

Explanation

Correct Answer

D:

A replay-safe manifest distinguishes outstanding items from completed or already submitted work and prevents accidental duplicate processing.

Incorrect Answers

A:

The deadline tolerates deferred execution and requests are independent, making batch orchestration a suitable operational pattern; the relevant precondition would be a nightly backlog consists of independent invoices due by the next morning. The question instead describes a batch worker may retry failed items after network interruption.

B:

A transaction depends on the output now; asynchronous batch scheduling would not satisfy the immediate completion dependency; the alternative would make sense if a cashier needs a tax calculation before finishing the same transaction. It does not directly fix the observed condition: a batch worker may retry failed items after network interruption.

C:

When the team wants to know whether extraction quality changes over several nightly runs, this would be useful because aggregate success status does not prove accuracy; scored examples reveal changes in extraction behavior over time; this item instead calls for a way to avoid creating duplicate submissions.

Question 227

A coordinator controls which tools each research worker may call. A drafting worker must produce prose using validated notes, not new unreviewed sources. What should be implemented at the appropriate boundary?

  1. Give the writer the validated evidence artifact and no external search tool
  2. Provide a computation tool and the curated input dataset to the analyst
  3. Let the coordinator select subsequent tools or specialists using completed status fields
  4. Allow approved retrieval tools while withholding document-writing capabilities

Explanation

Correct Answer

A:

Removing discovery tools prevents the writing stage from introducing fresh unvetted material after validation. This matches the workflow goal to restrict synthesis to approved evidence rather than introducing an unrelated action.

Incorrect Answers

B:

The analyst's task is arithmetic and interpretation, not unrestricted retrieval; limited tools reduce irrelevant calls and scope creep; the alternative would make sense if an analyst must compute statistics on previously collected evidence. It does not directly fix the observed condition: a drafting worker must produce prose using validated notes, not new unreviewed sources.

C:

This measure fits a different trigger: an orchestrator requires two worker outputs to choose the next branch. In that setting, A routing decision depends on multiple worker results; it belongs at the point that sees the combined state; it does not resolve the present need to restrict synthesis to approved evidence.

D:

Tool distribution should follow the worker's role; disallowing writes prevents discovery tasks from changing records; the relevant precondition would be a source-finding worker needs web search but cannot modify internal documents. The question instead describes a drafting worker must produce prose using validated notes, not new unreviewed sources.

Question 228

A repository team decides how to package MCP integrations for collaboration. The same server appears in multiple configuration layers with conflicting commands. Which option best meets the operational need?

  1. Keep the project-scoped server definition under version control without secrets
  2. Inspect the effective MCP configuration and resolve duplicate definitions
  3. Use a local-only MCP definition rather than committing it to shared project settings
  4. Review changed schemas and run integration checks before team rollout

Explanation

Correct Answer

B:

MCP discovery depends on active connection configuration; changing model instructions cannot fix an incorrect server command.

Incorrect Answers

A:

Versioned project configuration gives contributors consistent definitions while leaving credentials outside shared files; the relevant precondition would be a project-wide tool server has been approved for every contributor to that repository. The question instead describes the same server appears in multiple configuration layers with conflicting commands.

C:

Local scope prevents unreviewed capabilities from silently appearing in every contributor's workspace; the alternative would make sense if a developer tests an experimental local server that colleagues must not use. It does not directly fix the observed condition: the same server appears in multiple configuration layers with conflicting commands.

D:

This measure fits a different trigger: a server package is upgraded and its tool arguments change. In that setting, Server version changes can break tool calls even when connection setup succeeds, so schema compatibility requires testing; it does not resolve the present need to diagnose precedence instead of editing prompts.

Question 229

A team maintains path-aware architecture conventions in Claude Code. A new package directory does not match the existing rules. What is the most appropriate design choice?

  1. Revise path patterns and test their effective selection
  2. Give mobile and server trees distinct scoped guidance
  3. Keep shared guidance in one place and scoped exceptions explicitly narrow
  4. Attach the review procedure to infrastructure configuration patterns

Explanation

Correct Answer

A:

Rule files are not intelligent about renamed folders; matching must be maintained alongside structural refactors. The relevant requirement is to update the configuration when repository layout changes, and this measure operates at that decision point.

Incorrect Answers

B:

Relevant architecture instructions reduce accidental cross-project assumptions when a monorepo contains multiple build systems; the relevant precondition would be the mobile app follows a different package layout from the server. The question instead describes a new package directory does not match the existing rules.

C:

This measure fits a different trigger: a scoped rule repeats a contradictory general rule verbatim. In that setting, Redundant copies can drift; exceptions should state exactly what changes for a subset of files; it does not resolve the present need to update the configuration when repository layout changes.

D:

A path-aware rule gives extra caution where necessary without slowing unrelated text edits; the alternative would make sense if only infrastructure-as-code files require a plan-before-change policy. It does not directly fix the observed condition: a new package directory does not match the existing rules.

Question 230

A code assistant must choose among repository inspection and execution tools. A configuration change must be proposed but not written until review. Which response is most reliable?

  1. Allow file reads and searches while withholding file-edit tools
  2. Invoke the approved test runner and inspect its exit status and output
  3. Use repository search or symbol discovery on the relevant scope
  4. Give the scanner read access and leave release operations to the pipeline gate
  5. Restrict edits and shell working directory to the authorized path

Explanation

Correct Answer

A:

A planning phase should not silently mutate the workspace before the human approves the proposed patch. For the stated case, the check is whether the system can retain read-only capability during planning; this approach supplies that control.

Incorrect Answers

B:

A model-generated prediction cannot verify runtime behavior; the observed test result is the relevant evidence; the alternative would make sense if the team asks whether the current test suite passes after a patch. It does not directly fix the observed condition: a configuration change must be proposed but not written until review.

C:

Name-location tasks need index or text inspection; shell execution is unnecessary and may introduce side effects; the relevant precondition would be an engineer only asks which files define an interface by name. The question instead describes a configuration change must be proposed but not written until review.

D:

When a scanner output must be checked for known-risk entries before publishing, this would be useful because a worker that inspects risk should not control whether risky artifacts are published, especially before independent evaluation; this item instead calls for a way to retain read-only capability during planning.

E:

This measure fits a different trigger: a test-fix worker has permission to edit only a small package. In that setting, Worker-specific file boundaries reduce the risk that a narrowly assigned fix changes unrelated services; it does not resolve the present need to retain read-only capability during planning.

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