Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 64

Question 630

A team operating a customer account helpdesk agent wants to establish two separate behaviors: contradictory evidence is reviewed before action; in addition, a conclusion supported by model agreement is also tested against source evidence. Which plan implements both?

  1. select the most recent-looking sentence without checking its source; and promote agreement alone to a guaranteed correctness signal
  2. mark the inconsistency and request resolution or human verification; and require source and business-rule checks regardless of model agreement
  3. mark the inconsistency and request resolution or human verification; and promote agreement alone to a guaranteed correctness signal
  4. select the most recent-looking sentence without checking its source; and require source and business-rule checks regardless of model agreement

Explanation

Correct Answer

B:

In a customer account helpdesk agent, conflicting source claims are a knowledge gap, not authorization to guess; alongside this, correlated model errors can survive consensus without external verification, satisfying the distinct requirements that contradictory evidence is reviewed before action and that a conclusion supported by model agreement is also tested against source evidence.

Incorrect Answers

A:

In a customer account helpdesk agent, recency alone cannot establish which contradictory record is correct and shared biases can produce identical wrong conclusions; the design would fail both conditions: contradictory evidence is reviewed before action and a conclusion supported by model agreement is also tested against source evidence.

C:

In a customer account helpdesk agent, conflicting source claims are a knowledge gap, not authorization to guess, but shared biases can produce identical wrong conclusions; therefore the separate requirement that a conclusion supported by model agreement is also tested against source evidence would remain unsatisfied.

D:

In a customer account helpdesk agent, correlated model errors can survive consensus without external verification, but recency alone cannot establish which contradictory record is correct; therefore the separate requirement that contradictory evidence is reviewed before action would remain unsatisfied.

Question 631

The release gate for an analyst research coordinator requires that temporary service faults can be retried selectively; it also requires that a delegated agent receives the background it needs to act. Which TWO actions meet the specified checks? Choose TWO.

  1. retry every error with the same request indefinitely
  2. pass the necessary task context and references explicitly at invocation
  3. include a retryable classification and bounded retry advice
  4. assume the child agent sees all previous coordinator turns

Explanation

Correct Answers

B:

For an analyst research coordinator, subagents do not automatically inherit the coordinator's entire conversational history, which directly establishes that a delegated agent receives the background it needs to act; the separate requirement that temporary service faults can be retried selectively is handled by a different action.

C:

For an analyst research coordinator, typed transient failures allow backoff without treating permanent errors the same way, which directly establishes that temporary service faults can be retried selectively; the separate requirement that a delegated agent receives the background it needs to act is handled by a different action.

Incorrect Answers

A:

For an analyst research coordinator, permission and validation failures will not heal through repetition, so this option cannot establish that temporary service faults can be retried selectively; correctly addressing a delegated agent receives the background it needs to act elsewhere does not repair that deficiency.

D:

For an analyst research coordinator, implicit context inheritance makes delegation fragile and incomplete, so this option cannot establish that a delegated agent receives the background it needs to act; correctly addressing temporary service faults can be retried selectively elsewhere does not repair that deficiency.

Question 632

When designing an engineering knowledge assistant, the requirement is not satisfied unless a delegated agent receives the background it needs to act and two teams' analyses are reconciled before publication. Which implementation addresses the two independent conditions?

  1. pass the necessary task context and references explicitly at invocation; and publish whichever branch finishes last
  2. assume the child agent sees all previous coordinator turns; and add a synthesis stage with explicit conflict-resolution criteria
  3. pass the necessary task context and references explicitly at invocation; and add a synthesis stage with explicit conflict-resolution criteria
  4. implement only pass the necessary task context and references explicitly at invocation; make no corresponding change for the second requirement
  5. assume the child agent sees all previous coordinator turns; and publish whichever branch finishes last

Explanation

Correct Answer

C:

In an engineering knowledge assistant, subagents do not automatically inherit the coordinator's entire conversational history; alongside this, merging divergent work requires an authoritative step rather than choosing by arrival time, satisfying the distinct requirements that a delegated agent receives the background it needs to act and that two teams' analyses are reconciled before publication.

Incorrect Answers

A:

In an engineering knowledge assistant, subagents do not automatically inherit the coordinator's entire conversational history, but run order is not an evidence or quality standard; therefore the separate requirement that two teams' analyses are reconciled before publication would remain unsatisfied.

B:

In an engineering knowledge assistant, merging divergent work requires an authoritative step rather than choosing by arrival time, but implicit context inheritance makes delegation fragile and incomplete; therefore the separate requirement that a delegated agent receives the background it needs to act would remain unsatisfied.

D:

In an engineering knowledge assistant, subagents do not automatically inherit the coordinator's entire conversational history, but the independent mechanism for the other criterion is absent: merging divergent work requires an authoritative step rather than choosing by arrival time; as written, the design does not ensure that two teams' analyses are reconciled before publication.

E:

In an engineering knowledge assistant, implicit context inheritance makes delegation fragile and incomplete and run order is not an evidence or quality standard; the design would fail both conditions: a delegated agent receives the background it needs to act and two teams' analyses are reconciled before publication.

Question 633

Before expanding an order-status triage agent, the owner must demonstrate that resuming a session does not bypass present-day authorization, while also demonstrating that each tool interface has a single clear operation boundary. Which design provides both controls?

  1. treat old approval messages as unlimited authorization for later writes; and use a single universal tool that infers read versus write from prose
  2. recheck permissions and current resource state before new side effects; and separate read-only lookup from side-effecting update interfaces
  3. treat old approval messages as unlimited authorization for later writes; and separate read-only lookup from side-effecting update interfaces
  4. recheck permissions and current resource state before new side effects; and use a single universal tool that infers read versus write from prose

Explanation

Correct Answer

B:

In an order-status triage agent, conversation continuity is not a replacement for live access control or fresh tool state; alongside this, clear operation boundaries support least privilege and understandable approvals, satisfying the distinct requirements that resuming a session does not bypass present-day authorization and that each tool interface has a single clear operation boundary.

Incorrect Answers

A:

In an order-status triage agent, permissions and operational context can change between sessions and overloading hides effects and makes execution review harder; the design would fail both conditions: resuming a session does not bypass present-day authorization and each tool interface has a single clear operation boundary.

C:

In an order-status triage agent, clear operation boundaries support least privilege and understandable approvals, but permissions and operational context can change between sessions; therefore the separate requirement that resuming a session does not bypass present-day authorization would remain unsatisfied.

D:

In an order-status triage agent, conversation continuity is not a replacement for live access control or fresh tool state, 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.

Question 634

Before expanding a research PDF metadata pipeline, the owner must demonstrate that a broad tool catalog does not overwhelm a small task, while also demonstrating that partial completion of a multi-step operation is handled safely. Which design provides both controls?

  1. supply only context-relevant tools for the current agent; and track individual stage outcomes and use explicit recovery or compensating actions
  2. load all available MCP tools into every specialist turn; and track individual stage outcomes and use explicit recovery or compensating actions
  3. load all available MCP tools into every specialist turn; and treat the first successful tool call as proof that the entire workflow succeeded
  4. supply only context-relevant tools for the current agent; and treat the first successful tool call as proof that the entire workflow succeeded

Explanation

Correct Answer

A:

In a research PDF metadata pipeline, reducing irrelevant tool definitions improves selection and saves context; alongside this, an orchestrated workflow must not report overall success while a required step failed, satisfying the distinct requirements that a broad tool catalog does not overwhelm a small task and that partial completion of a multi-step operation is handled safely.

Incorrect Answers

B:

In a research PDF metadata pipeline, an orchestrated workflow must not report overall success while a required step failed, but tool bloat makes the available actions harder to discriminate; therefore the separate requirement that a broad tool catalog does not overwhelm a small task would remain unsatisfied.

C:

In a research PDF metadata pipeline, tool bloat makes the available actions harder to discriminate and later actions can fail after an earlier side effect; the design would fail both conditions: a broad tool catalog does not overwhelm a small task and partial completion of a multi-step operation is handled safely.

D:

In a research PDF metadata pipeline, reducing irrelevant tool definitions improves selection and saves context, 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.

Question 635

Tests for an insurance document extraction agent will pass only if each tool interface has a single clear operation boundary and if persistent ambiguous data receives human review. Which proposed design directly satisfies both checks?

  1. implement only separate read-only lookup from side-effecting update interfaces; make no corresponding change for the second requirement
  2. separate read-only lookup from side-effecting update interfaces; and stop after defined retries and route unresolved cases to a reviewer
  3. separate read-only lookup from side-effecting update interfaces; and keep retrying until the model finally emits something that passes
  4. use a single universal tool that infers read versus write from prose; and keep retrying until the model finally emits something that passes
  5. use a single universal tool that infers read versus write from prose; and stop after defined retries and route unresolved cases to a reviewer

Explanation

Correct Answer

B:

In an insurance document extraction agent, clear operation boundaries support least privilege and understandable approvals; alongside this, a bounded escalation path prevents infinite loops and invented resolutions, satisfying the distinct requirements that each tool interface has a single clear operation boundary and that persistent ambiguous data receives human review.

Incorrect Answers

A:

In an insurance document extraction agent, clear operation boundaries support least privilege and understandable approvals, but the independent mechanism for the other criterion is absent: a bounded escalation path prevents infinite loops and invented resolutions; as written, the design does not ensure that persistent ambiguous data receives human review.

C:

In an insurance document extraction agent, clear operation boundaries support least privilege and understandable approvals, but unlimited retries waste resources and can reinforce guesses; therefore the separate requirement that persistent ambiguous data receives human review would remain unsatisfied.

D:

In an insurance document extraction agent, overloading hides effects and makes execution review harder and unlimited retries waste resources and can reinforce guesses; the design would fail both conditions: each tool interface has a single clear operation boundary and persistent ambiguous data receives human review.

E:

In an insurance document extraction agent, a bounded escalation path prevents infinite loops and invented resolutions, 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.

Question 636

In an automated CI review service, the immediate design requirement is that ambiguous requirements are resolved before a destructive migration. Which change most directly establishes that behavior?

  1. exit the planning stage and perform scoped implementation with verification
  2. clarify scope and acceptance criteria while planning
  3. use plan mode to inspect dependencies and propose steps before execution
  4. make assumptions and migrate production files immediately

Explanation

Correct Answer

B:

For an automated CI review service, a reviewed plan should resolve unknown targets and rollback expectations before writes; this directly establishes that ambiguous requirements are resolved before a destructive migration instead of relying on an unrelated workflow control.

Incorrect Answers

A:

Although planning describes intended operations while execution applies and checks them, that only supports the distinct goal that an approved plan is translated into controlled edits and tests, not the requirement that ambiguous requirements are resolved before a destructive migration in an automated CI review service.

C:

Although planning separates investigation and approval from code mutation, that only supports the distinct goal that a risky refactor is evaluated before any files change, not the requirement that ambiguous requirements are resolved before a destructive migration in an automated CI review service.

D:

In an automated CI review service, guessed intent can create costly irreversible changes; as a result, the stated requirement that ambiguous requirements are resolved before a destructive migration would still not be met.

Question 637

In a billing-dispute support agent, the acceptance criteria require that persistent ambiguous data receives human review and that a function call accepts only valid argument shapes. Which paired implementation meets both?

  1. keep retrying until the model finally emits something that passes; and define the tool input schema, enable strict tool use where available, and validate inputs in the handler
  2. keep retrying until the model finally emits something that passes; and allow arbitrary keys and silently coerce incorrect types
  3. stop after defined retries and route unresolved cases to a reviewer; and allow arbitrary keys and silently coerce incorrect types
  4. stop after defined retries and route unresolved cases to a reviewer; and define the tool input schema, enable strict tool use where available, and validate inputs in the handler

Explanation

Correct Answer

D:

In a billing-dispute support agent, a bounded escalation path prevents infinite loops and invented resolutions; alongside this, strict tool inputs constrain generation, and handler validation provides a final enforceable argument boundary, satisfying the distinct requirements that persistent ambiguous data receives human review and that a function call accepts only valid argument shapes.

Incorrect Answers

A:

In a billing-dispute support agent, strict tool inputs constrain generation, and handler validation provides a final enforceable argument boundary, but unlimited retries waste resources and can reinforce guesses; therefore the separate requirement that persistent ambiguous data receives human review would remain unsatisfied.

B:

In a billing-dispute support agent, unlimited retries waste resources and can reinforce guesses and unvalidated coercion can change intended operations or silently drop important fields; the design would fail both conditions: persistent ambiguous data receives human review and a function call accepts only valid argument shapes.

C:

In a billing-dispute support agent, a bounded escalation path prevents infinite loops and invented resolutions, but unvalidated coercion can change intended operations or silently drop important fields; therefore the separate requirement that a function call accepts only valid argument shapes would remain unsatisfied.

Question 638

An incident postmortem for a literature synthesis workflow requires controls so that a truncated response is not mistaken for successful completion and that multiple subagent outputs can be traced to their individual tasks. Which paired remedy covers both failure modes?

  1. handle max_tokens as truncation rather than end_turn; and attach task IDs and source references to returned artifacts
  2. handle max_tokens as truncation rather than end_turn; and strip origin metadata from each response before merging
  3. interpret every non-tool stop reason as a successful final answer; and strip origin metadata from each response before merging
  4. interpret every non-tool stop reason as a successful final answer; and attach task IDs and source references to returned artifacts
  5. implement only handle max_tokens as truncation rather than end_turn; make no corresponding change for the second requirement

Explanation

Correct Answer

A:

In a literature synthesis workflow, reaching an output limit is a stop condition but does not establish task completion; alongside this, traceable results let the coordinator identify which delegation produced each claim, satisfying the distinct requirements that a truncated response is not mistaken for successful completion and that multiple subagent outputs can be traced to their individual tasks.

Incorrect Answers

B:

In a literature synthesis workflow, reaching an output limit is a stop condition but does not establish task completion, 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 literature synthesis workflow, a length-limited result may be incomplete or unusable and source-less aggregation impairs diagnosis and fact checking; the design would fail both conditions: a truncated response is not mistaken for successful completion and multiple subagent outputs can be traced to their individual tasks.

D:

In a literature synthesis workflow, traceable results let the coordinator identify which delegation produced each claim, but a length-limited result may be incomplete or unusable; therefore the separate requirement that a truncated response is not mistaken for successful completion would remain unsatisfied.

E:

In a literature synthesis workflow, reaching an output limit is a stop condition but does not establish task completion, but the independent mechanism for the other criterion is absent: traceable results let the coordinator identify which delegation produced each claim; as written, the design does not ensure that multiple subagent outputs can be traced to their individual tasks.

Question 639

A team operating a typed API generator project wants to establish two separate behaviors: an approved plan is translated into controlled edits and tests; in addition, two nested packages follow distinct framework conventions. Which plan implements both?

  1. exit the planning stage and perform scoped implementation with verification; and write one ambiguous instruction for both packages
  2. implement only exit the planning stage and perform scoped implementation with verification; make no corresponding change for the second requirement
  3. remain in read-only planning and claim the code was updated; and create separate path-specific rule files for each package
  4. remain in read-only planning and claim the code was updated; and write one ambiguous instruction for both packages
  5. exit the planning stage and perform scoped implementation with verification; and create separate path-specific rule files for each package

Explanation

Correct Answer

E:

In a typed API generator project, planning describes intended operations while execution applies and checks them; alongside this, scoped conventions avoid contradictory patterns between codebases sharing a monorepo, satisfying the distinct requirements that an approved plan is translated into controlled edits and tests and that two nested packages follow distinct framework conventions.

Incorrect Answers

A:

In a typed API generator project, planning describes intended operations while execution applies and checks them, but the model must guess which framework convention applies; therefore the separate requirement that two nested packages follow distinct framework conventions would remain unsatisfied.

B:

In a typed API generator project, planning describes intended operations while execution applies and checks them, but the independent mechanism for the other criterion is absent: scoped conventions avoid contradictory patterns between codebases sharing a monorepo; as written, the design does not ensure that two nested packages follow distinct framework conventions.

C:

In a typed API generator project, scoped conventions avoid contradictory patterns between codebases sharing a monorepo, but a proposal alone does not modify repository state; therefore the separate requirement that an approved plan is translated into controlled edits and tests would remain unsatisfied.

D:

In a typed API generator project, a proposal alone does not modify repository state and the model must guess which framework convention applies; the design would fail both conditions: an approved plan is translated into controlled edits and tests and two nested packages follow distinct framework conventions.

Question 640

Production readiness for a developer codebase exploration assistant depends on ensuring that two nested packages follow distinct framework conventions and that only a nested service uses specialized implementation rules. Select the most appropriate combined approach.

  1. write one ambiguous instruction for both packages; and place a dedicated CLAUDE.md in the relevant directory
  2. write one ambiguous instruction for both packages; and put all service-only details in the repository root instruction file
  3. create separate path-specific rule files for each package; and put all service-only details in the repository root instruction file
  4. create separate path-specific rule files for each package; and place a dedicated CLAUDE.md in the relevant directory

Explanation

Correct Answer

D:

In a developer codebase exploration assistant, scoped conventions avoid contradictory patterns between codebases sharing a monorepo; alongside this, directory-specific context can refine instructions for work under that subtree, satisfying the distinct requirements that two nested packages follow distinct framework conventions and that only a nested service uses specialized implementation rules.

Incorrect Answers

A:

In a developer codebase exploration assistant, directory-specific context can refine instructions for work under that subtree, but the model must guess which framework convention applies; therefore the separate requirement that two nested packages follow distinct framework conventions would remain unsatisfied.

B:

In a developer codebase exploration assistant, the model must guess which framework convention applies and globalizing narrow rules adds irrelevant instructions to unrelated work; the design would fail both conditions: two nested packages follow distinct framework conventions and only a nested service uses specialized implementation rules.

C:

In a developer codebase exploration assistant, scoped conventions avoid contradictory patterns between codebases sharing a monorepo, 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.

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