Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 91

Question 900

Tests for a code policy conformance job will pass only if partial completion of a multi-step operation is handled safely and if secrets are not embedded in a shared MCP setup. Which proposed design directly satisfies both checks?

  1. implement only inject credentials through approved environment or secret-management mechanisms; make no corresponding change for the first requirement
  2. track individual stage outcomes and use explicit recovery or compensating actions; and inject credentials through approved environment or secret-management mechanisms
  3. treat the first successful tool call as proof that the entire workflow succeeded; and place long-lived access tokens directly in .mcp.json
  4. treat the first successful tool call as proof that the entire workflow succeeded; and inject credentials through approved environment or secret-management mechanisms
  5. track individual stage outcomes and use explicit recovery or compensating actions; and place long-lived access tokens directly in .mcp.json

Explanation

Correct Answer

B:

In a code policy conformance job, an orchestrated workflow must not report overall success while a required step failed; alongside this, separating credentials from versioned configuration limits accidental disclosure, satisfying the distinct requirements that partial completion of a multi-step operation is handled safely and that secrets are not embedded in a shared MCP setup.

Incorrect Answers

A:

In a code policy conformance job, separating credentials from versioned configuration limits accidental disclosure, but the independent mechanism for the other criterion is absent: an orchestrated workflow must not report overall success while a required step failed; as written, the design does not ensure that partial completion of a multi-step operation is handled safely.

C:

In a code policy conformance job, later actions can fail after an earlier side effect and repository history can expose tokens to everyone with code access; the design would fail both conditions: partial completion of a multi-step operation is handled safely and secrets are not embedded in a shared MCP setup.

D:

In a code policy conformance job, separating credentials from versioned configuration limits accidental disclosure, but later actions can fail after an earlier side effect; therefore the separate requirement that partial completion of a multi-step operation is handled safely would remain unsatisfied.

E:

In a code policy conformance job, an orchestrated workflow must not report overall success while a required step failed, but repository history can expose tokens to everyone with code access; therefore the separate requirement that secrets are not embedded in a shared MCP setup would remain unsatisfied.

Question 901

In a pre-merge code inspection workflow, the design must ensure that a team compares two candidate patches against the same baseline and that CI failures provide actionable evidence to maintainers. Choose TWO independent changes that directly meet these two requirements.

  1. discard the runner logs and keep only the assistant's final sentence
  2. persist structured errors, relevant logs, and a bounded diagnostic summary
  3. evaluate alternatives using the same tests and scoped diffs
  4. choose whichever patch has the more persuasive natural-language summary

Explanation

Correct Answers

B:

For a pre-merge code inspection workflow, stable artifacts help maintainers reproduce the problem without rerunning a chat, which directly establishes that CI failures provide actionable evidence to maintainers; the separate requirement that a team compares two candidate patches against the same baseline is handled by a different action.

C:

For a pre-merge code inspection workflow, consistent acceptance evidence reveals which candidate improves the relevant behavior, which directly establishes that a team compares two candidate patches against the same baseline; the separate requirement that CI failures provide actionable evidence to maintainers is handled by a different action.

Incorrect Answers

A:

For a pre-merge code inspection workflow, the missing trace impedes diagnosis of command or infrastructure failures, so this option cannot establish that CI failures provide actionable evidence to maintainers; correctly addressing a team compares two candidate patches against the same baseline elsewhere does not repair that deficiency.

D:

For a pre-merge code inspection workflow, rhetorical fluency does not establish code correctness, so this option cannot establish that a team compares two candidate patches against the same baseline; correctly addressing CI failures provide actionable evidence to maintainers elsewhere does not repair that deficiency.

Question 902

For a service dependency exploration tool, a reviewer identifies two independent obligations: a large investigation spans unrelated subsystems and tool-call logging does not modify a successful tool result. Which proposal should be approved?

  1. implement only delegate bounded subsystem searches and merge path-based findings; make no corresponding change for the second requirement
  2. delegate bounded subsystem searches and merge path-based findings; and rewrite every tool result as a free-form paraphrase
  3. delegate bounded subsystem searches and merge path-based findings; and use an observation-only hook for telemetry when rewriting is unnecessary
  4. give every specialist the complete monorepo transcript; and use an observation-only hook for telemetry when rewriting is unnecessary
  5. give every specialist the complete monorepo transcript; and rewrite every tool result as a free-form paraphrase

Explanation

Correct Answer

C:

In a service dependency exploration tool, subsystem partitioning permits parallel work without duplicating full repository context; alongside this, separating instrumentation from mutation reduces the risk of corrupting evidence, satisfying the distinct requirements that a large investigation spans unrelated subsystems and that tool-call logging does not modify a successful tool result.

Incorrect Answers

A:

In a service dependency exploration tool, subsystem partitioning permits parallel work without duplicating full repository context, but the independent mechanism for the other criterion is absent: separating instrumentation from mutation reduces the risk of corrupting evidence; as written, the design does not ensure that tool-call logging does not modify a successful tool result.

B:

In a service dependency exploration tool, subsystem partitioning permits parallel work without duplicating full repository context, but summaries can discard fields that subsequent steps require; therefore the separate requirement that tool-call logging does not modify a successful tool result would remain unsatisfied.

D:

In a service dependency exploration tool, separating instrumentation from mutation reduces the risk of corrupting evidence, but each branch pays the same large context cost and increases noise; therefore the separate requirement that a large investigation spans unrelated subsystems would remain unsatisfied.

E:

In a service dependency exploration tool, each branch pays the same large context cost and increases noise and summaries can discard fields that subsequent steps require; the design would fail both conditions: a large investigation spans unrelated subsystems and tool-call logging does not modify a successful tool result.

Question 903

Production readiness for a pull-request review pipeline depends on ensuring that policy decisions use measurable conditions rather than tone and that a model learns a boundary between superficially similar cases. Select the most appropriate combined approach.

  1. describe exact observable triggers and exception rules; and show only easy positive samples
  2. ask the model to use its general intuition about severity; and include contrasting positive and negative demonstrations
  3. describe exact observable triggers and exception rules; and include contrasting positive and negative demonstrations
  4. ask the model to use its general intuition about severity; and show only easy positive samples
  5. implement only describe exact observable triggers and exception rules; make no corresponding change for the second requirement

Explanation

Correct Answer

C:

In a pull-request review pipeline, operationalized criteria reduce disagreement across similar inputs; alongside this, near-boundary examples communicate the decisive distinction and reduce false positives, satisfying the distinct requirements that policy decisions use measurable conditions rather than tone and that a model learns a boundary between superficially similar cases.

Incorrect Answers

A:

In a pull-request review pipeline, operationalized criteria reduce disagreement across similar inputs, but the model may generalize the label to unrelated but similar-looking items; therefore the separate requirement that a model learns a boundary between superficially similar cases would remain unsatisfied.

B:

In a pull-request review pipeline, near-boundary examples communicate the decisive distinction and reduce false positives, but unquantified intuition varies between cases and reviewers; therefore the separate requirement that policy decisions use measurable conditions rather than tone would remain unsatisfied.

D:

In a pull-request review pipeline, unquantified intuition varies between cases and reviewers and the model may generalize the label to unrelated but similar-looking items; the design would fail both conditions: policy decisions use measurable conditions rather than tone and a model learns a boundary between superficially similar cases.

E:

In a pull-request review pipeline, operationalized criteria reduce disagreement across similar inputs, but the independent mechanism for the other criterion is absent: near-boundary examples communicate the decisive distinction and reduce false positives; as written, the design does not ensure that a model learns a boundary between superficially similar cases.

Question 904

An audit of a shipping-exception helpdesk identifies two unmet requirements: the model can use a tool result in its next decision and a missing MCP capability is diagnosed before changing prompts. Select TWO actions that remedy those particular deficiencies.

  1. put the tool output only in application logs
  2. increase the system prompt length to compensate for disconnected tools
  3. verify server connection, transport, and exposed tool availability
  4. append the assistant tool-use response and matching user tool_result blocks to history

Explanation

Correct Answers

C:

For a shipping-exception helpdesk, tool discovery depends on a live configured integration, not just the prompt wording, which directly establishes that a missing MCP capability is diagnosed before changing prompts; the separate requirement that the model can use a tool result in its next decision is handled by a different action.

D:

For a shipping-exception helpdesk, Claude needs the prior request and the execution output in the conversation for the next inference, which directly establishes that the model can use a tool result in its next decision; the separate requirement that a missing MCP capability is diagnosed before changing prompts is handled by a different action.

Incorrect Answers

A:

For a shipping-exception helpdesk, unseen logs cannot inform the model's next decision, so this option cannot establish that the model can use a tool result in its next decision; correctly addressing a missing MCP capability is diagnosed before changing prompts elsewhere does not repair that deficiency.

B:

For a shipping-exception helpdesk, a model cannot call a capability absent from its available tool registry, so this option cannot establish that a missing MCP capability is diagnosed before changing prompts; correctly addressing the model can use a tool result in its next decision elsewhere does not repair that deficiency.

Question 905

The release checklist for a large-codebase architecture explorer explicitly checks that conflicting subagent findings produce one accountable interpretation; it also checks that multiple subagent outputs can be traced to their individual tasks. What coordinated solution meets the checklist?

  1. implement only attach task IDs and source references to returned artifacts; make no corresponding change for the first requirement
  2. have the coordinator compare claims and sources before synthesis; and strip origin metadata from each response before merging
  3. let the fastest subagent override other findings automatically; and attach task IDs and source references to returned artifacts
  4. have the coordinator compare claims and sources before synthesis; and attach task IDs and source references to returned artifacts
  5. let the fastest subagent override other findings automatically; and strip origin metadata from each response before merging

Explanation

Correct Answer

D:

In a large-codebase architecture explorer, central reconciliation can distinguish evidence disagreement from a formatting difference; alongside this, traceable results let the coordinator identify which delegation produced each claim, satisfying the distinct requirements that conflicting subagent findings produce one accountable interpretation and that multiple subagent outputs can be traced to their individual tasks.

Incorrect Answers

A:

In a large-codebase architecture explorer, traceable results let the coordinator identify which delegation produced each claim, but the independent mechanism for the other criterion is absent: central reconciliation can distinguish evidence disagreement from a formatting difference; as written, the design does not ensure that conflicting subagent findings produce one accountable interpretation.

B:

In a large-codebase architecture explorer, central reconciliation can distinguish evidence disagreement from a formatting difference, but source-less aggregation impairs diagnosis and fact checking; therefore the separate requirement that multiple subagent outputs can be traced to their individual tasks would remain unsatisfied.

C:

In a large-codebase architecture explorer, traceable results let the coordinator identify which delegation produced each claim, but completion order is not a sound evidence-ranking rule; therefore the separate requirement that conflicting subagent findings produce one accountable interpretation would remain unsatisfied.

E:

In a large-codebase architecture explorer, completion order is not a sound evidence-ranking rule and source-less aggregation impairs diagnosis and fact checking; the design would fail both conditions: conflicting subagent findings produce one accountable interpretation and multiple subagent outputs can be traced to their individual tasks.

Question 906

For a technical due-diligence report generator, a reviewer identifies two independent obligations: the model can distinguish two similar business tools and a downstream subagent acts on concrete upstream results. Which proposal should be approved?

  1. give each tool a specific purpose and exclusion examples in its description; and send only a generic instruction to continue the work
  2. give both tools vague names with identical generic descriptions; and pass the preceding artifact identifier and a focused summary to the next agent
  3. implement only give each tool a specific purpose and exclusion examples in its description; make no corresponding change for the second requirement
  4. give both tools vague names with identical generic descriptions; and send only a generic instruction to continue the work
  5. give each tool a specific purpose and exclusion examples in its description; and pass the preceding artifact identifier and a focused summary to the next agent

Explanation

Correct Answer

E:

In a technical due-diligence report generator, discriminative descriptions reduce wrong routing when operations sound alike; alongside this, explicit handoff preserves the specific result that the dependent stage must process, satisfying the distinct requirements that the model can distinguish two similar business tools and that a downstream subagent acts on concrete upstream results.

Incorrect Answers

A:

In a technical due-diligence report generator, discriminative descriptions reduce wrong routing when operations sound alike, but the later agent cannot know which evidence or state was produced previously; therefore the separate requirement that a downstream subagent acts on concrete upstream results would remain unsatisfied.

B:

In a technical due-diligence report generator, explicit handoff preserves the specific result that the dependent stage must process, but a model cannot reliably select on boundaries that the interface never explains; therefore the separate requirement that the model can distinguish two similar business tools would remain unsatisfied.

C:

In a technical due-diligence report generator, discriminative descriptions reduce wrong routing when operations sound alike, but the independent mechanism for the other criterion is absent: explicit handoff preserves the specific result that the dependent stage must process; as written, the design does not ensure that a downstream subagent acts on concrete upstream results.

D:

In a technical due-diligence report generator, a model cannot reliably select on boundaries that the interface never explains and the later agent cannot know which evidence or state was produced previously; the design would fail both conditions: the model can distinguish two similar business tools and a downstream subagent acts on concrete upstream results.

Question 907

An audit of a nightly code quality job identifies two unmet requirements: CI failures provide actionable evidence to maintainers and API conventions apply only to source files under an API directory. Select TWO actions that remedy those particular deficiencies.

  1. add a .claude/rules file with paths matching the API subtree
  2. discard the runner logs and keep only the assistant's final sentence
  3. persist structured errors, relevant logs, and a bounded diagnostic summary
  4. put API-specific formatting rules in always-on project guidance

Explanation

Correct Answers

A:

For a nightly code quality job, path-scoped rules load when the target context matches, keeping unrelated tasks clean, which directly establishes that API conventions apply only to source files under an API directory; the separate requirement that CI failures provide actionable evidence to maintainers is handled by a different action.

C:

For a nightly code quality job, stable artifacts help maintainers reproduce the problem without rerunning a chat, which directly establishes that CI failures provide actionable evidence to maintainers; the separate requirement that API conventions apply only to source files under an API directory is handled by a different action.

Incorrect Answers

B:

For a nightly code quality job, the missing trace impedes diagnosis of command or infrastructure failures, so this option cannot establish that CI failures provide actionable evidence to maintainers; correctly addressing API conventions apply only to source files under an API directory elsewhere does not repair that deficiency.

D:

For a nightly code quality job, frontend and documentation tasks then receive irrelevant instructions, so this option cannot establish that API conventions apply only to source files under an API directory; correctly addressing CI failures provide actionable evidence to maintainers elsewhere does not repair that deficiency.

Question 908

Before expanding a customs paperwork processor, the owner must demonstrate that several requested client tools are returned in one response, while also demonstrating that repeated delivery of one action request cannot create duplicate writes. Which design provides both controls?

  1. execute only the first returned tool call; and blindly resubmit the same write after every timeout
  2. execute every tool_use block and return a result with each corresponding ID; and blindly resubmit the same write after every timeout
  3. execute every tool_use block and return a result with each corresponding ID; and use an idempotency key and confirm the prior commit before retrying
  4. execute only the first returned tool call; and use an idempotency key and confirm the prior commit before retrying

Explanation

Correct Answer

C:

In a customs paperwork processor, a multi-tool assistant turn can request independent actions and must receive all matching outcomes; alongside this, side-effecting operations need deduplication against timeouts and repeated messages, satisfying the distinct requirements that several requested client tools are returned in one response and that repeated delivery of one action request cannot create duplicate writes.

Incorrect Answers

A:

In a customs paperwork processor, the unanswered tool uses remain unresolved and later reasoning loses evidence and the first attempt might have committed despite a missing acknowledgment; the design would fail both conditions: several requested client tools are returned in one response and repeated delivery of one action request cannot create duplicate writes.

B:

In a customs paperwork processor, a multi-tool assistant turn can request independent actions and must receive all matching outcomes, but the first attempt might have committed despite a missing acknowledgment; therefore the separate requirement that repeated delivery of one action request cannot create duplicate writes would remain unsatisfied.

D:

In a customs paperwork processor, side-effecting operations need deduplication against timeouts and repeated messages, but the unanswered tool uses remain unresolved and later reasoning loses evidence; therefore the separate requirement that several requested client tools are returned in one response would remain unsatisfied.

Question 909

For a change-impact analysis service, a reviewer identifies two independent obligations: long project guidance is divided into maintainable source files and secrets are not embedded in a shared MCP setup. Which proposal should be approved?

  1. duplicate full instruction paragraphs in every directory; and place long-lived access tokens directly in .mcp.json
  2. use CLAUDE.md imports to organize referenced instructions; and place long-lived access tokens directly in .mcp.json
  3. use CLAUDE.md imports to organize referenced instructions; and inject credentials through approved environment or secret-management mechanisms
  4. duplicate full instruction paragraphs in every directory; and inject credentials through approved environment or secret-management mechanisms

Explanation

Correct Answer

C:

In a change-impact analysis service, modular imports keep a short entry point while preserving reviewed details in separate files; alongside this, separating credentials from versioned configuration limits accidental disclosure, satisfying the distinct requirements that long project guidance is divided into maintainable source files and that secrets are not embedded in a shared MCP setup.

Incorrect Answers

A:

In a change-impact analysis service, copying guidance encourages drift when policies change and repository history can expose tokens to everyone with code access; the design would fail both conditions: long project guidance is divided into maintainable source files and secrets are not embedded in a shared MCP setup.

B:

In a change-impact analysis service, modular imports keep a short entry point while preserving reviewed details in separate files, but repository history can expose tokens to everyone with code access; therefore the separate requirement that secrets are not embedded in a shared MCP setup would remain unsatisfied.

D:

In a change-impact analysis service, separating credentials from versioned configuration limits accidental disclosure, but copying guidance encourages drift when policies change; therefore the separate requirement that long project guidance is divided into maintainable source files would remain unsatisfied.

Question 910

An architecture review of a customer entitlement assistant finds two missing guarantees: a missing value is represented without fabricating a string; conflicting sources are not silently flattened into one answer. What two-part change closes both gaps?

  1. allow null in the field schema and define its semantics; and choose one assertion and omit the contradiction without investigation
  2. force the model to populate every unknown value with a guess; and choose one assertion and omit the contradiction without investigation
  3. force the model to populate every unknown value with a guess; and report the disagreement and explain evidence quality or recency
  4. implement only allow null in the field schema and define its semantics; make no corresponding change for the second requirement
  5. allow null in the field schema and define its semantics; and report the disagreement and explain evidence quality or recency

Explanation

Correct Answer

E:

In a customer entitlement assistant, explicit nullable types distinguish genuine absence from empty or invented data; alongside this, explicit conflicts preserve uncertainty rather than manufacturing unanimity, satisfying the distinct requirements that a missing value is represented without fabricating a string and that conflicting sources are not silently flattened into one answer.

Incorrect Answers

A:

In a customer entitlement assistant, explicit nullable types distinguish genuine absence from empty or invented data, but suppressed disagreement creates false confidence; therefore the separate requirement that conflicting sources are not silently flattened into one answer would remain unsatisfied.

B:

In a customer entitlement assistant, fabricated values preserve schema appearance but destroy factual accuracy and suppressed disagreement creates false confidence; the design would fail both conditions: a missing value is represented without fabricating a string and conflicting sources are not silently flattened into one answer.

C:

In a customer entitlement assistant, explicit conflicts preserve uncertainty rather than manufacturing unanimity, but fabricated values preserve schema appearance but destroy factual accuracy; therefore the separate requirement that a missing value is represented without fabricating a string would remain unsatisfied.

D:

In a customer entitlement assistant, explicit nullable types distinguish genuine absence from empty or invented data, but the independent mechanism for the other criterion is absent: explicit conflicts preserve uncertainty rather than manufacturing unanimity; as written, the design does not ensure that conflicting sources are not silently flattened into one answer.

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