Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 87

Question 860

In a technical review of a service-credit approval assistant, the architect must address the following constraints together: a changed account record supersedes an old conversational value; an extraction error is caught before storage. Which change set is most defensible?

  1. reuse the value from the first turn without checking for change; and validate parsed data against application rules and source constraints
  2. refresh the authoritative source before a consequential write; and validate parsed data against application rules and source constraints
  3. refresh the authoritative source before a consequential write; and write the response to the database without checking required invariants
  4. implement only refresh the authoritative source before a consequential write; make no corresponding change for the second requirement
  5. reuse the value from the first turn without checking for change; and write the response to the database without checking required invariants

Explanation

Correct Answer

B:

In a service-credit approval assistant, current external state protects against stale facts preserved in context; alongside this, local validation detects missing cross-field relationships even when the JSON is syntactically valid, satisfying the distinct requirements that a changed account record supersedes an old conversational value and that an extraction error is caught before storage.

Incorrect Answers

A:

In a service-credit approval assistant, local validation detects missing cross-field relationships even when the JSON is syntactically valid, but conversations do not automatically reflect later database updates; therefore the separate requirement that a changed account record supersedes an old conversational value would remain unsatisfied.

C:

In a service-credit approval assistant, current external state protects against stale facts preserved in context, but valid-looking output may still violate business rules; therefore the separate requirement that an extraction error is caught before storage would remain unsatisfied.

D:

In a service-credit approval assistant, current external state protects against stale facts preserved in context, but the independent mechanism for the other criterion is absent: local validation detects missing cross-field relationships even when the JSON is syntactically valid; as written, the design does not ensure that an extraction error is caught before storage.

E:

In a service-credit approval assistant, conversations do not automatically reflect later database updates and valid-looking output may still violate business rules; the design would fail both conditions: a changed account record supersedes an old conversational value and an extraction error is caught before storage.

Question 861

An audit of a security-advisory aggregation workflow asks how to make sure that delegated investigations stay within bounded scope. Which implementation should be selected?

  1. ask every subagent to investigate everything it can find
  2. specify the question, allowed sources, expected output, and stopping criterion
  3. define each subagent role and expose the minimum matching tools
  4. spawn independent specialists in the same delegation stage

Explanation

Correct Answer

B:

For a security-advisory aggregation workflow, a task contract prevents specialist exploration from swallowing the coordinator's larger goal; this directly establishes that delegated investigations stay within bounded scope instead of relying on an unrelated workflow control.

Incorrect Answers

A:

In a security-advisory aggregation workflow, unbounded exploration dilutes signal and inflates context and latency; as a result, the stated requirement that delegated investigations stay within bounded scope would still not be met.

C:

Although tool scoping shrinks both decision confusion and the blast radius of delegated actions, that only supports the distinct goal that specialized agents work with only relevant permissions, not the requirement that delegated investigations stay within bounded scope in a security-advisory aggregation workflow.

D:

Although concurrent subagents can examine non-overlapping modules before the coordinator merges their reports, that only supports the distinct goal that separate code modules can be inspected in parallel, not the requirement that delegated investigations stay within bounded scope in a security-advisory aggregation workflow.

Question 862

For a competitive benchmarking pipeline, a reviewer identifies two independent obligations: an alternate implementation is explored without rewriting the original branch and independent stages of a complex job run efficiently. Which proposal should be approved?

  1. continue changing the original session as though it were an isolated experiment; and divide work along independent outputs and run those stages in parallel
  2. fork the existing session into a separate line of exploration; and divide work along independent outputs and run those stages in parallel
  3. continue changing the original session as though it were an isolated experiment; and split the job into an arbitrary fixed chain of agents
  4. fork the existing session into a separate line of exploration; and split the job into an arbitrary fixed chain of agents

Explanation

Correct Answer

B:

In a competitive benchmarking pipeline, a fork supports competing paths while preserving the original session history; alongside this, dependency-aware decomposition avoids serial work when branches do not consume one another, satisfying the distinct requirements that an alternate implementation is explored without rewriting the original branch and that independent stages of a complex job run efficiently.

Incorrect Answers

A:

In a competitive benchmarking pipeline, dependency-aware decomposition avoids serial work when branches do not consume one another, but the experiment then shares and alters the active thread's direction; therefore the separate requirement that an alternate implementation is explored without rewriting the original branch would remain unsatisfied.

C:

In a competitive benchmarking pipeline, the experiment then shares and alters the active thread's direction and a rigid sequence creates waits even where no dependency exists; the design would fail both conditions: an alternate implementation is explored without rewriting the original branch and independent stages of a complex job run efficiently.

D:

In a competitive benchmarking pipeline, a fork supports competing paths while preserving the original session history, but a rigid sequence creates waits even where no dependency exists; therefore the separate requirement that independent stages of a complex job run efficiently would remain unsatisfied.

Question 863

The release checklist for a monorepo feature modernization project explicitly checks that a required validation step runs before the final response even when the model does not support forced tool choice; it also checks that a destructive tool receives human authorization at execution time. What coordinated solution meets the checklist?

  1. rely on a prompt asking the model to call the validator; and interrupt or deny the tool until an authorization callback approves it
  2. rely on a prompt asking the model to call the validator; and collect a general approval at the start of the conversation
  3. implement only interrupt or deny the tool until an authorization callback approves it; make no corresponding change for the first requirement
  4. enforce the validator in application control flow, using forced tool_choice only with compatible models; and collect a general approval at the start of the conversation
  5. enforce the validator in application control flow, using forced tool_choice only with compatible models; and interrupt or deny the tool until an authorization callback approves it

Explanation

Correct Answer

E:

In a monorepo feature modernization project, application-level orchestration guarantees the required stage across models, while forced tool_choice is limited to compatible models and settings; alongside this, execution gates make the approval binding on the selected action and inputs, satisfying the distinct requirements that a required validation step runs before the final response even when the model does not support forced tool choice and that a destructive tool receives human authorization at execution time.

Incorrect Answers

A:

In a monorepo feature modernization project, execution gates make the approval binding on the selected action and inputs, but prompt instructions do not guarantee execution of the validator or override model-specific tool-choice restrictions; therefore the separate requirement that a required validation step runs before the final response even when the model does not support forced tool choice would remain unsatisfied.

B:

In a monorepo feature modernization project, prompt instructions do not guarantee execution of the validator or override model-specific tool-choice restrictions and blanket consent does not verify the later high-risk operation; the design would fail both conditions: a required validation step runs before the final response even when the model does not support forced tool choice and a destructive tool receives human authorization at execution time.

C:

In a monorepo feature modernization project, execution gates make the approval binding on the selected action and inputs, but the independent mechanism for the other criterion is absent: application-level orchestration guarantees the required stage across models, while forced tool_choice is limited to compatible models and settings; as written, the design does not ensure that a required validation step runs before the final response even when the model does not support forced tool choice.

D:

In a monorepo feature modernization project, application-level orchestration guarantees the required stage across models, while forced tool_choice is limited to compatible models and settings, but blanket consent does not verify the later high-risk operation; therefore the separate requirement that a destructive tool receives human authorization at execution time would remain unsatisfied.

Question 864

When designing a supplier invoice extractor, the requirement is not satisfied unless few-shot material does not contain contradictory answer labels and source identity survives multi-agent research handoffs. Which implementation addresses the two independent conditions?

  1. keep historical examples even after the labeling policy changes; and strip all citation fields before sending the writer its inputs
  2. keep historical examples even after the labeling policy changes; and carry citation identifiers and retrieval context through synthesis
  3. audit examples against the current decision rubric before use; and strip all citation fields before sending the writer its inputs
  4. audit examples against the current decision rubric before use; and carry citation identifiers and retrieval context through synthesis

Explanation

Correct Answer

D:

In a supplier invoice extractor, consistent labels reinforce the intended criteria rather than training ambiguity into the prompt; alongside this, provenance metadata must remain attached when intermediate summaries are merged, satisfying the distinct requirements that few-shot material does not contain contradictory answer labels and that source identity survives multi-agent research handoffs.

Incorrect Answers

A:

In a supplier invoice extractor, stale demonstrations can contradict the updated standard and the writer cannot reliably reconstruct which source supports each conclusion; the design would fail both conditions: few-shot material does not contain contradictory answer labels and source identity survives multi-agent research handoffs.

B:

In a supplier invoice extractor, provenance metadata must remain attached when intermediate summaries are merged, but stale demonstrations can contradict the updated standard; therefore the separate requirement that few-shot material does not contain contradictory answer labels would remain unsatisfied.

C:

In a supplier invoice extractor, consistent labels reinforce the intended criteria rather than training ambiguity into the prompt, but the writer cannot reliably reconstruct which source supports each conclusion; therefore the separate requirement that source identity survives multi-agent research handoffs would remain unsatisfied.

Question 865

During review of a warranty claims assistant, an unmet criterion states that a changed account record supersedes an old conversational value. Which measure resolves this specific gap?

  1. pass a compact structured record of facts, decisions, pending tasks, and sources
  2. track pending approvals and required next steps in durable task state
  3. refresh the authoritative source before a consequential write
  4. reuse the value from the first turn without checking for change

Explanation

Correct Answer

C:

For a warranty claims assistant, current external state protects against stale facts preserved in context; this directly establishes that a changed account record supersedes an old conversational value instead of relying on an unrelated workflow control.

Incorrect Answers

A:

Although focused handoffs reduce context size while protecting continuity-critical details, that only supports the distinct goal that a handoff retains only action-relevant context, not the requirement that a changed account record supersedes an old conversational value in a warranty claims assistant.

B:

Although persisted obligations survive when natural-language history is shortened, that only supports the distinct goal that context reduction does not erase unresolved obligations, not the requirement that a changed account record supersedes an old conversational value in a warranty claims assistant.

D:

In a warranty claims assistant, conversations do not automatically reflect later database updates; as a result, the stated requirement that a changed account record supersedes an old conversational value would still not be met.

Question 866

Before expanding a legacy service decomposition, the owner must demonstrate that a source change is validated against the right dependency chain, while also demonstrating that the repository assistant discovers filenames matching a pattern. Which design provides both controls?

  1. review only the edited line with no caller context; and use Glob to locate candidate paths
  2. trace imports and call sites from the touched module outward; and search entire file contents with Grep to find paths by naming pattern
  3. review only the edited line with no caller context; and search entire file contents with Grep to find paths by naming pattern
  4. trace imports and call sites from the touched module outward; and use Glob to locate candidate paths

Explanation

Correct Answer

D:

In a legacy service decomposition, dependency-focused exploration spots affected behavior without indiscriminate reading; alongside this, path globbing finds files by names and directory structure before content inspection, satisfying the distinct requirements that a source change is validated against the right dependency chain and that the repository assistant discovers filenames matching a pattern.

Incorrect Answers

A:

In a legacy service decomposition, path globbing finds files by names and directory structure before content inspection, but local correctness may miss consumers that rely on the old behavior; therefore the separate requirement that a source change is validated against the right dependency chain would remain unsatisfied.

B:

In a legacy service decomposition, dependency-focused exploration spots affected behavior without indiscriminate reading, but content search answers a different question and may miss empty files; therefore the separate requirement that the repository assistant discovers filenames matching a pattern would remain unsatisfied.

C:

In a legacy service decomposition, local correctness may miss consumers that rely on the old behavior and content search answers a different question and may miss empty files; the design would fail both conditions: a source change is validated against the right dependency chain and the repository assistant discovers filenames matching a pattern.

Question 867

In a backend API maintenance team, the immediate design requirement is that the repository can review and version its custom workflows. Which change most directly establishes that behavior?

  1. store project skills in the repository's .claude/skills structure
  2. distribute the canonical skill only through private direct messages
  3. create a documented reusable skill in SKILL.md
  4. define how the skill accepts and uses invocation arguments

Explanation

Correct Answer

A:

For a backend API maintenance team, version-controlled skills evolve alongside tests and conventions with code review; this directly establishes that the repository can review and version its custom workflows instead of relying on an unrelated workflow control.

Incorrect Answers

B:

In a backend API maintenance team, unversioned copies lose the team's approved definition; as a result, the stated requirement that the repository can review and version its custom workflows would still not be met.

C:

Although skills package instructions and supporting resources for invocation across similar tasks, that only supports the distinct goal that a repeatable code-review procedure is available as a named workflow, not the requirement that the repository can review and version its custom workflows in a backend API maintenance team.

D:

Although parameterized workflows apply one reviewed procedure to different resources without hardcoding targets, that only supports the distinct goal that a shared skill can take a supplied target name or file path, not the requirement that the repository can review and version its custom workflows in a backend API maintenance team.

Question 868

Tests for a warranty claims assistant will pass only if invalid parameters lead to a clear correction rather than a blind retry and if the model can use a tool result in its next decision. Which proposed design directly satisfies both checks?

  1. hide the invalid field behind a generic internal-error message; and append the assistant tool-use response and matching user tool_result blocks to history
  2. return a non-retryable validation error with the field and expected constraint; and put the tool output only in application logs
  3. hide the invalid field behind a generic internal-error message; and put the tool output only in application logs
  4. return a non-retryable validation error with the field and expected constraint; and append the assistant tool-use response and matching user tool_result blocks to history

Explanation

Correct Answer

D:

In a warranty claims assistant, field-specific feedback lets the caller fix the input before resubmission; alongside this, Claude needs the prior request and the execution output in the conversation for the next inference, satisfying the distinct requirements that invalid parameters lead to a clear correction rather than a blind retry and that the model can use a tool result in its next decision.

Incorrect Answers

A:

In a warranty claims assistant, Claude needs the prior request and the execution output in the conversation for the next inference, but the client lacks the information needed to repair the request; therefore the separate requirement that invalid parameters lead to a clear correction rather than a blind retry would remain unsatisfied.

B:

In a warranty claims assistant, field-specific feedback lets the caller fix the input before resubmission, but unseen logs cannot inform the model's next decision; therefore the separate requirement that the model can use a tool result in its next decision would remain unsatisfied.

C:

In a warranty claims assistant, the client lacks the information needed to repair the request and unseen logs cannot inform the model's next decision; the design would fail both conditions: invalid parameters lead to a clear correction rather than a blind retry and the model can use a tool result in its next decision.

Question 869

An incident postmortem for an automated CI review service requires controls so that developers can explicitly invoke an established project procedure and that the implementation plan receives stakeholder sign-off. Which paired remedy covers both failure modes?

  1. write undocumented one-off ad hoc prompts for each engineer; and present ordered changes, risks, and verification steps before switching to execution
  2. make the skill available under a recognizable slash invocation; and treat the agent's internally generated plan as implicit approval
  3. make the skill available under a recognizable slash invocation; and present ordered changes, risks, and verification steps before switching to execution
  4. write undocumented one-off ad hoc prompts for each engineer; and treat the agent's internally generated plan as implicit approval

Explanation

Correct Answer

C:

In an automated CI review service, a discoverable workflow gives the team a consistent entry point for repetitive work; alongside this, an explicit review point aligns the agent's work with the authorized scope, satisfying the distinct requirements that developers can explicitly invoke an established project procedure and that the implementation plan receives stakeholder sign-off.

Incorrect Answers

A:

In an automated CI review service, an explicit review point aligns the agent's work with the authorized scope, but unshared prompts make results hard to reproduce and review; therefore the separate requirement that developers can explicitly invoke an established project procedure would remain unsatisfied.

B:

In an automated CI review service, a discoverable workflow gives the team a consistent entry point for repetitive work, but the system cannot infer external authorization from its own reasoning; therefore the separate requirement that the implementation plan receives stakeholder sign-off would remain unsatisfied.

D:

In an automated CI review service, unshared prompts make results hard to reproduce and review and the system cannot infer external authorization from its own reasoning; the design would fail both conditions: developers can explicitly invoke an established project procedure and the implementation plan receives stakeholder sign-off.

Question 870

An architecture review of a structured case-file ingestion workflow finds two missing guarantees: source identity survives multi-agent research handoffs; a long workflow preserves confirmed facts separately from speculation. What two-part change closes both gaps?

  1. strip all citation fields before sending the writer its inputs; and combine guesses and verified attributes in one unqualified summary
  2. strip all citation fields before sending the writer its inputs; and mark verified, user-reported, and unknown state explicitly
  3. carry citation identifiers and retrieval context through synthesis; and combine guesses and verified attributes in one unqualified summary
  4. carry citation identifiers and retrieval context through synthesis; and mark verified, user-reported, and unknown state explicitly
  5. implement only carry citation identifiers and retrieval context through synthesis; make no corresponding change for the second requirement

Explanation

Correct Answer

D:

In a structured case-file ingestion workflow, provenance metadata must remain attached when intermediate summaries are merged; alongside this, epistemic labels stop a tentative suggestion becoming a confirmed action parameter, satisfying the distinct requirements that source identity survives multi-agent research handoffs and that a long workflow preserves confirmed facts separately from speculation.

Incorrect Answers

A:

In a structured case-file ingestion workflow, the writer cannot reliably reconstruct which source supports each conclusion and downstream operations may rely on an assumption as though it were checked; the design would fail both conditions: source identity survives multi-agent research handoffs and a long workflow preserves confirmed facts separately from speculation.

B:

In a structured case-file ingestion workflow, epistemic labels stop a tentative suggestion becoming a confirmed action parameter, but the writer cannot reliably reconstruct which source supports each conclusion; therefore the separate requirement that source identity survives multi-agent research handoffs would remain unsatisfied.

C:

In a structured case-file ingestion workflow, provenance metadata must remain attached when intermediate summaries are merged, 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.

E:

In a structured case-file ingestion workflow, provenance metadata must remain attached when intermediate summaries are merged, but the independent mechanism for the other criterion is absent: epistemic labels stop a tentative suggestion becoming a confirmed action parameter; as written, the design does not ensure that a long workflow preserves confirmed facts separately from speculation.

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