Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 99

Question 980

In a technical review of a contract-analysis research network, the architect must address the following constraints together: a complex document receives focused quality checks; a consequential claim receives independent human sign-off. Which change set is most defensible?

  1. use one generic request to make the document better; and accept model confidence as a substitute for authorization
  2. use one generic request to make the document better; and route high-impact or low-evidence results to an authorized reviewer
  3. separate factual, structural, and policy review passes; and accept model confidence as a substitute for authorization
  4. separate factual, structural, and policy review passes; and route high-impact or low-evidence results to an authorized reviewer

Explanation

Correct Answer

D:

In a contract-analysis research network, specialized criteria can detect different failure modes without one overloaded evaluation instruction; alongside this, human approval supplies accountability where a mistaken model result has material cost, satisfying the distinct requirements that a complex document receives focused quality checks and that a consequential claim receives independent human sign-off.

Incorrect Answers

A:

In a contract-analysis research network, an undifferentiated review can overlook specific required constraints and a model score does not confer business or legal authority; the design would fail both conditions: a complex document receives focused quality checks and a consequential claim receives independent human sign-off.

B:

In a contract-analysis research network, human approval supplies accountability where a mistaken model result has material cost, but an undifferentiated review can overlook specific required constraints; therefore the separate requirement that a complex document receives focused quality checks would remain unsatisfied.

C:

In a contract-analysis research network, specialized criteria can detect different failure modes without one overloaded evaluation instruction, but a model score does not confer business or legal authority; therefore the separate requirement that a consequential claim receives independent human sign-off would remain unsatisfied.

Question 981

Production readiness for a structured report conversion service depends on ensuring that extracted records are labeled only when required fields agree and that a partially failed batch is repaired efficiently. Select the most appropriate combined approach.

  1. state the conjunction of required matching conditions; and identify failed records and resubmit only safe retry candidates
  2. classify as a match when any one keyword overlaps; and rerun the entire completed batch whenever one record fails
  3. classify as a match when any one keyword overlaps; and identify failed records and resubmit only safe retry candidates
  4. state the conjunction of required matching conditions; and rerun the entire completed batch whenever one record fails

Explanation

Correct Answer

A:

In a structured report conversion service, explicit multi-field criteria prevent a single suggestive field from dominating the outcome; alongside this, selective retry avoids paying again for already successful items, satisfying the distinct requirements that extracted records are labeled only when required fields agree and that a partially failed batch is repaired efficiently.

Incorrect Answers

B:

In a structured report conversion service, a shared keyword does not prove record identity and successful work would be duplicated unnecessarily; the design would fail both conditions: extracted records are labeled only when required fields agree and a partially failed batch is repaired efficiently.

C:

In a structured report conversion service, selective retry avoids paying again for already successful items, but a shared keyword does not prove record identity; therefore the separate requirement that extracted records are labeled only when required fields agree would remain unsatisfied.

D:

In a structured report conversion service, explicit multi-field criteria prevent a single suggestive field from dominating the outcome, but successful work would be duplicated unnecessarily; therefore the separate requirement that a partially failed batch is repaired efficiently would remain unsatisfied.

Question 982

In a multi-repository troubleshooting agent, the acceptance criteria require that each tool interface has a single clear operation boundary and that only a nested service uses specialized implementation rules. Which paired implementation meets both?

  1. separate read-only lookup from side-effecting update interfaces; and place a dedicated CLAUDE.md in the relevant directory
  2. implement only separate read-only lookup from side-effecting update interfaces; make no corresponding change for the second requirement
  3. separate read-only lookup from side-effecting update interfaces; and put all service-only details in the repository root instruction file
  4. use a single universal tool that infers read versus write from prose; and place a dedicated CLAUDE.md in the relevant directory
  5. use a single universal tool that infers read versus write from prose; and put all service-only details in the repository root instruction file

Explanation

Correct Answer

A:

In a multi-repository troubleshooting agent, clear operation boundaries support least privilege and understandable approvals; alongside this, directory-specific context can refine instructions for work under that subtree, satisfying the distinct requirements that each tool interface has a single clear operation boundary and that only a nested service uses specialized implementation rules.

Incorrect Answers

B:

In a multi-repository troubleshooting agent, clear operation boundaries support least privilege and understandable approvals, but the independent mechanism for the other criterion is absent: directory-specific context can refine instructions for work under that subtree; as written, the design does not ensure that only a nested service uses specialized implementation rules.

C:

In a multi-repository troubleshooting agent, clear operation boundaries support least privilege and understandable approvals, but globalizing narrow rules adds irrelevant instructions to unrelated work; therefore the separate requirement that only a nested service uses specialized implementation rules would remain unsatisfied.

D:

In a multi-repository troubleshooting agent, directory-specific context can refine instructions for work under that subtree, but overloading hides effects and makes execution review harder; therefore the separate requirement that each tool interface has a single clear operation boundary would remain unsatisfied.

E:

In a multi-repository troubleshooting agent, overloading hides effects and makes execution review harder and globalizing narrow rules adds irrelevant instructions to unrelated work; the design would fail both conditions: each tool interface has a single clear operation boundary and only a nested service uses specialized implementation rules.

Question 983

In a source-grounded policy synthesis service, the design must ensure that a model learns a boundary between superficially similar cases and that a consequential claim receives independent human sign-off. Choose TWO independent changes that directly meet these two requirements.

  1. route high-impact or low-evidence results to an authorized reviewer
  2. accept model confidence as a substitute for authorization
  3. include contrasting positive and negative demonstrations
  4. show only easy positive samples

Explanation

Correct Answers

A:

For a source-grounded policy synthesis service, human approval supplies accountability where a mistaken model result has material cost, which directly establishes that a consequential claim receives independent human sign-off; the separate requirement that a model learns a boundary between superficially similar cases is handled by a different action.

C:

For a source-grounded policy synthesis service, near-boundary examples communicate the decisive distinction and reduce false positives, which directly establishes that a model learns a boundary between superficially similar cases; the separate requirement that a consequential claim receives independent human sign-off is handled by a different action.

Incorrect Answers

B:

For a source-grounded policy synthesis service, a model score does not confer business or legal authority, so this option cannot establish that a consequential claim receives independent human sign-off; correctly addressing a model learns a boundary between superficially similar cases elsewhere does not repair that deficiency.

D:

For a source-grounded policy synthesis service, the model may generalize the label to unrelated but similar-looking items, so this option cannot establish that a model learns a boundary between superficially similar cases; correctly addressing a consequential claim receives independent human sign-off elsewhere does not repair that deficiency.

Question 984

For an automated test-generation workflow, acceptance depends on demonstrating that a high-risk result passes deterministic checks in addition to model critique. Which approach is the strongest fit?

  1. combine independent review with schema and business-rule validation
  2. give the reviewer the source and rubric, not only the draft's self-rating
  3. replace all validation checks with a second model's opinion
  4. separate factual, structural, and policy review passes

Explanation

Correct Answer

A:

For an automated test-generation workflow, a critique may improve quality but cannot enforce machine-testable invariants by itself; this directly establishes that a high-risk result passes deterministic checks in addition to model critique instead of relying on an unrelated workflow control.

Incorrect Answers

B:

Although independent evidence lets a critic challenge unsupported assertions, that only supports the distinct goal that the reviewer does not simply echo the generator's confidence, not the requirement that a high-risk result passes deterministic checks in addition to model critique in an automated test-generation workflow.

C:

In an automated test-generation workflow, probabilistic agreement does not guarantee compliance; as a result, the stated requirement that a high-risk result passes deterministic checks in addition to model critique would still not be met.

D:

Although specialized criteria can detect different failure modes without one overloaded evaluation instruction, that only supports the distinct goal that a complex document receives focused quality checks, not the requirement that a high-risk result passes deterministic checks in addition to model critique in an automated test-generation workflow.

Question 985

Tests for a billing-dispute support agent will pass only if a long workflow preserves confirmed facts separately from speculation and if confidence scores reflect available evidence and uncertainty. Which proposed design directly satisfies both checks?

  1. mark verified, user-reported, and unknown state explicitly; and set confidence to high whenever the explanation is detailed
  2. mark verified, user-reported, and unknown state explicitly; and calibrate confidence against known outcomes and source support
  3. combine guesses and verified attributes in one unqualified summary; and calibrate confidence against known outcomes and source support
  4. combine guesses and verified attributes in one unqualified summary; and set confidence to high whenever the explanation is detailed

Explanation

Correct Answer

B:

In a billing-dispute support agent, epistemic labels stop a tentative suggestion becoming a confirmed action parameter; alongside this, observed error rates and evidence quality are better than a fluent self-rating, satisfying the distinct requirements that a long workflow preserves confirmed facts separately from speculation and that confidence scores reflect available evidence and uncertainty.

Incorrect Answers

A:

In a billing-dispute support agent, epistemic labels stop a tentative suggestion becoming a confirmed action parameter, but verbal detail is not a validated reliability estimate; therefore the separate requirement that confidence scores reflect available evidence and uncertainty would remain unsatisfied.

C:

In a billing-dispute support agent, observed error rates and evidence quality are better than a fluent self-rating, but downstream operations may rely on an assumption as though it were checked; therefore the separate requirement that a long workflow preserves confirmed facts separately from speculation would remain unsatisfied.

D:

In a billing-dispute support agent, downstream operations may rely on an assumption as though it were checked and verbal detail is not a validated reliability estimate; the design would fail both conditions: a long workflow preserves confirmed facts separately from speculation and confidence scores reflect available evidence and uncertainty.

Question 986

While preparing a competitive benchmarking pipeline for deployment, engineers must ensure that the system distinguishes acceptable and unacceptable edge cases; they must also ensure that the reviewer does not simply echo the generator's confidence. Which combination best satisfies those requirements?

  1. specify inclusion criteria and explicit exclusion conditions; and provide the generator's score as the sole review criterion
  2. specify inclusion criteria and explicit exclusion conditions; and give the reviewer the source and rubric, not only the draft's self-rating
  3. provide only a broad topic label as the classification rule; and provide the generator's score as the sole review criterion
  4. provide only a broad topic label as the classification rule; and give the reviewer the source and rubric, not only the draft's self-rating

Explanation

Correct Answer

B:

In a competitive benchmarking pipeline, negative boundaries teach the model when a near-match must not trigger; alongside this, independent evidence lets a critic challenge unsupported assertions, satisfying the distinct requirements that the system distinguishes acceptable and unacceptable edge cases and that the reviewer does not simply echo the generator's confidence.

Incorrect Answers

A:

In a competitive benchmarking pipeline, negative boundaries teach the model when a near-match must not trigger, but anchoring on a self-assigned grade hides actual defects; therefore the separate requirement that the reviewer does not simply echo the generator's confidence would remain unsatisfied.

C:

In a competitive benchmarking pipeline, generic labels leave ambiguous boundary cases undefined and anchoring on a self-assigned grade hides actual defects; the design would fail both conditions: the system distinguishes acceptable and unacceptable edge cases and the reviewer does not simply echo the generator's confidence.

D:

In a competitive benchmarking pipeline, independent evidence lets a critic challenge unsupported assertions, but generic labels leave ambiguous boundary cases undefined; therefore the separate requirement that the system distinguishes acceptable and unacceptable edge cases would remain unsatisfied.

Question 987

Before expanding a returns-policy assistant, the owner must demonstrate that schema-valid output is also checked against source evidence, while also demonstrating that a failed retrieval agent does not masquerade as having found no matches. Which design provides both controls?

  1. assume a schema-conforming response is factually correct; and return a typed failure status with its origin to the coordinator
  2. assume a schema-conforming response is factually correct; and convert every retrieval failure into an empty list
  3. validate semantic constraints and provenance in application logic; and convert every retrieval failure into an empty list
  4. validate semantic constraints and provenance in application logic; and return a typed failure status with its origin to the coordinator

Explanation

Correct Answer

D:

In a returns-policy assistant, JSON shape does not prove that an extracted claim is supported by the document; alongside this, an operational error must remain distinguishable from a valid empty result, satisfying the distinct requirements that schema-valid output is also checked against source evidence and that a failed retrieval agent does not masquerade as having found no matches.

Incorrect Answers

A:

In a returns-policy assistant, an operational error must remain distinguishable from a valid empty result, but syntactic conformance cannot guarantee source-level truth; therefore the separate requirement that schema-valid output is also checked against source evidence would remain unsatisfied.

B:

In a returns-policy assistant, syntactic conformance cannot guarantee source-level truth and the final report could incorrectly claim no sources exist; the design would fail both conditions: schema-valid output is also checked against source evidence and a failed retrieval agent does not masquerade as having found no matches.

C:

In a returns-policy assistant, JSON shape does not prove that an extracted claim is supported by the document, but the final report could incorrectly claim no sources exist; therefore the separate requirement that a failed retrieval agent does not masquerade as having found no matches would remain unsatisfied.

Question 988

When designing an insurance document extraction agent, the requirement is not satisfied unless a classification uses a limited approved category set and the coordinator can identify which subagent produced a broken artifact. Which implementation addresses the two independent conditions?

  1. define precise output labels and mapping rules; and replace all nested errors with a generic coordinator failure
  2. define precise output labels and mapping rules; and propagate task IDs, error codes, and correlation references
  3. allow a free-form category for each case; and propagate task IDs, error codes, and correlation references
  4. allow a free-form category for each case; and replace all nested errors with a generic coordinator failure

Explanation

Correct Answer

B:

In an insurance document extraction agent, controlled categories reduce schema drift and make downstream analytics comparable; alongside this, tracing execution lineage helps route the incident to the right failing component, satisfying the distinct requirements that a classification uses a limited approved category set and that the coordinator can identify which subagent produced a broken artifact.

Incorrect Answers

A:

In an insurance document extraction agent, controlled categories reduce schema drift and make downstream analytics comparable, but the original failing operation becomes opaque; therefore the separate requirement that the coordinator can identify which subagent produced a broken artifact would remain unsatisfied.

C:

In an insurance document extraction agent, tracing execution lineage helps route the incident to the right failing component, but unbounded labels fragment the intended taxonomy; therefore the separate requirement that a classification uses a limited approved category set would remain unsatisfied.

D:

In an insurance document extraction agent, unbounded labels fragment the intended taxonomy and the original failing operation becomes opaque; the design would fail both conditions: a classification uses a limited approved category set and the coordinator can identify which subagent produced a broken artifact.

Question 989

During review of a multi-source policy brief service, an unmet criterion states that dependent analysis begins only when source retrieval is complete. Which measure resolves this specific gap?

  1. dispatch the independent branches concurrently through the coordinator
  2. record per-branch status and let the coordinator choose retry or partial response
  3. start the analysis and retrieval agents together without shared results
  4. run retrieval before analysis and pass its validated artifacts

Explanation

Correct Answer

D:

For a multi-source policy brief service, a downstream agent needs actual upstream findings rather than an assumed future result; this directly establishes that dependent analysis begins only when source retrieval is complete instead of relying on an unrelated workflow control.

Incorrect Answers

A:

Although parallel delegation reduces latency when individual investigations do not consume each other's outputs, that only supports the distinct goal that independent research branches can finish without waiting on one another, not the requirement that dependent analysis begins only when source retrieval is complete in a multi-source policy brief service.

B:

Although the orchestration layer must distinguish missing evidence from a negative factual finding, that only supports the distinct goal that the system can continue when one specialist returns an explicit failure, not the requirement that dependent analysis begins only when source retrieval is complete in a multi-source policy brief service.

C:

In a multi-source policy brief service, the analysis branch cannot evaluate evidence it has not received; as a result, the stated requirement that dependent analysis begins only when source retrieval is complete would still not be met.

Question 990

While preparing a repository search copilot for deployment, engineers must ensure that a large research job combines specialized findings into one answer; they must also ensure that a team shares a repository-specific MCP server configuration. Which combination best satisfies those requirements?

  1. have all subagents independently publish the final answer; and commit a reviewed project .mcp.json without credentials
  2. have all subagents independently publish the final answer; and configure each contributor's personal MCP registry manually
  3. implement only use a coordinator that delegates focused tasks and owns final synthesis; make no corresponding change for the second requirement
  4. use a coordinator that delegates focused tasks and owns final synthesis; and configure each contributor's personal MCP registry manually
  5. use a coordinator that delegates focused tasks and owns final synthesis; and commit a reviewed project .mcp.json without credentials

Explanation

Correct Answer

E:

In a repository search copilot, the hub can integrate distinct subagent outputs and resolve overlap without peer-to-peer chatter; alongside this, project-scoped configuration can be distributed with the repository while keeping secrets external, satisfying the distinct requirements that a large research job combines specialized findings into one answer and that a team shares a repository-specific MCP server configuration.

Incorrect Answers

A:

In a repository search copilot, project-scoped configuration can be distributed with the repository while keeping secrets external, but independent final outputs offer no single point to reconcile contradictions; therefore the separate requirement that a large research job combines specialized findings into one answer would remain unsatisfied.

B:

In a repository search copilot, independent final outputs offer no single point to reconcile contradictions and local personal settings do not create a reproducible team project setup; the design would fail both conditions: a large research job combines specialized findings into one answer and a team shares a repository-specific MCP server configuration.

C:

In a repository search copilot, the hub can integrate distinct subagent outputs and resolve overlap without peer-to-peer chatter, but the independent mechanism for the other criterion is absent: project-scoped configuration can be distributed with the repository while keeping secrets external; as written, the design does not ensure that a team shares a repository-specific MCP server configuration.

D:

In a repository search copilot, the hub can integrate distinct subagent outputs and resolve overlap without peer-to-peer chatter, but local personal settings do not create a reproducible team project setup; therefore the separate requirement that a team shares a repository-specific MCP server configuration would remain unsatisfied.

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