Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 4

Question 30

A coordinator has three worker results but no rule for selecting the final recommendation. What is the missing architectural component?

  1. Partition independent expert analyses and reserve integration for the orchestrator
  2. Send trivial deterministic steps directly and reserve specialists for nontrivial reasoning
  3. Add a deterministic aggregation policy for worker output acceptance
  4. Focus on holding publication until validation reports are reconciled with the extracted artifacts

Explanation

Correct Answer

C:

A merge contract can specify completeness, conflict handling and evidence precedence. Without it, the final recommendation may depend on incidental completion order.

Incorrect Answers

A:

Focused workers can receive targeted instructions and smaller context windows; its benefits for matching specialist agents to distinct competencies do not resolve the need for defining a reproducible worker merge policy.

B:

Agent delegation has token, latency and observability costs; supporting avoiding needless specialist agent invocations cannot replace the required defining a reproducible worker merge policy.

D:

Completion of one branch does not imply total workflow completion; that serves awaiting validation across distributed workers, whereas the case requires defining a reproducible worker merge policy.

Question 31

A remote system returns partial results before one shard times out. How should the result be represented?

  1. Provide a tool result marked isError for the failed operation
  2. Re-run time-sensitive reads after recovery and replace stale failure state
  3. Flag partial completion explicitly and identify which data remains unavailable
  4. Provide an explicit authorization failure rather than a misleading not-found response

Explanation

Correct Answer

C:

A partial response must not masquerade as complete evidence. Structured completeness metadata lets agents qualify claims or request targeted recovery.

Incorrect Answers

A:

For signaling execution failure in an MCP result, mCP distinguishes a tool execution error from a normal successful result with an error flag; however, this requirement concerns communicating incomplete tool output.

B:

Cached failure artifacts are not live evidence; that may serve refreshing state after a transient service failure, but it is no substitute for communicating incomplete tool output.

D:

Permission denials differ from nonexistent resources; the stated problem requires communicating incomplete tool output rather than distinguishing access denial from missing data.

Question 32

A subagent creates a multi-page analysis while the parent only needs a risk level and supporting citations. How should delegation be tuned?

  1. Limit both delegated instructions and available data connectors to the assigned case
  2. Retain approval for destructive decisions with the coordinator or an authorized human
  3. Specify a concise result schema suited to the parent's merge operation
  4. Insist on evidence of which documents and versions the worker actually consulted
  5. Launch a reviewer with an isolated prompt and the source requirements only

Explanation

Correct Answer

C:

An output shape with decision, evidence and uncertainty avoids spending context on unusable narration. Parent orchestration becomes easier when worker responses are fit for purpose.

Incorrect Answers

A:

Context minimization and access controls complement each other; the option would serve narrowing a worker's data-access boundary, not minimizing unnecessary worker output context as specified in the stem.

B:

Delegating writing does not imply delegating business authority; this addresses retaining authority for irreversible actions, while the question is about minimizing unnecessary worker output context.

D:

Although source and version metadata distinguish inspected facts from memory-based guesses, the choice addresses validating the evidentiary basis of a subagent answer and not the requested minimizing unnecessary worker output context.

E:

A clean review context reduces anchoring on the first draft; the deciding condition is minimizing unnecessary worker output context, whereas this action concerns maintaining independence for critique.

Question 33

The inventory server responds with a temporary overload error. What error guidance supports resilient clients?

  1. Surface a retryable temporary-failure classification with backoff guidance
  2. Include a trace or correlation identifier to the diagnostic result
  3. Categorize the business constraint as nontransient and require a different next step
  4. Emit an explicit authorization failure rather than a misleading not-found response

Explanation

Correct Answer

A:

Transient capacity problems may clear without changing the business request. The caller should apply bounded retries and jitter rather than treat the condition as a permanent absence of data.

Incorrect Answers

B:

Correlation metadata links user-facing failure to server logs and specific attempts; the unsatisfied requirement is identifying safely retryable service overload, not correlating errors across distributed tool services.

C:

Funds or policy failures generally need user or system-state changes, not immediate identical retries; in this scenario, avoiding futile retries on business-rule rejection is secondary to identifying safely retryable service overload.

D:

When considering distinguishing access denial from missing data, permission denials differ from nonexistent resources; the requirement for identifying safely retryable service overload remains unmet.

Question 34

A cross-service release has strict rollback and sequencing requirements. What planning information is indispensable?

  1. Map dependencies and produce a change plan before mutation
  2. Enforce the reviewed patch within its stated scope and verify the security regression
  3. Show the plan for approval before allowing write-capable execution
  4. Specify deployment order, checkpoints and rollback triggers before execution

Explanation

Correct Answer

D:

Multi-service updates have dependencies and failure domains that cannot be improvised safely. A planned sequence gives operators explicit hold and recovery points.

Incorrect Answers

A:

Discovery identifies relevant modules, tests and side effects; that addresses mapping dependencies before modifying unfamiliar code, whereas this case calls for preparing release sequencing and rollback.

B:

Replanning a settled narrow remediation can delay correction and introduce scope drift; this concerns avoiding needless replanning of an approved patch, leaving the requirement for preparing release sequencing and rollback outstanding.

C:

Sensitive migrations benefit from separating design review from mutation; requiring plan approval for a migration and preparing release sequencing and rollback are distinct responsibilities under these constraints.

Question 35

An MCP customer tool advertises that it returns a record but does not define what missing results look like. What should be added?

  1. Record the supported reason codes as a constrained enumeration
  2. Provide a narrow parameterized order-status operation instead of general SQL
  3. Emit structured, bounded results with pagination and relevant fields
  4. Document explicit success, not-found and error result shapes

Explanation

Correct Answer

D:

A clear output contract lets downstream agents tell absent records from service faults. Mixing missing data with errors can lead to unauthorized guesses or unnecessary retries.

Incorrect Answers

A:

Enumerated fields constrain inputs to a documented business vocabulary and support deterministic validation; this contributes to limiting parameters to valid business values without meeting the need for specifying the semantics of empty tool results.

B:

A purpose-limited interface minimizes injection and unauthorized data access; that serves constraining a tool to least-privilege business operations, whereas the case requires specifying the semantics of empty tool results.

C:

Because typed and paginated results let the agent request evidence progressively, the measure supports controlling the volume of tool output rather than the needed specifying the semantics of empty tool results.

Question 36

A developer wants reproducible outputs from a contract-risk assessor. What prompt asset should accompany the decision rubric?

  1. Declare evaluation priorities such as latency, correctness, budget and operational ownership
  2. List the exact fields, types and normalization rules required downstream
  3. Supply a clear output structure with explicit labels for evidence and risk rationale
  4. Define required positive findings, limitations, risks and outstanding decisions

Explanation

Correct Answer

C:

Combining an evaluation rubric with an output contract makes responses comparable. A free-form essay hides which claims support each rating.

Incorrect Answers

A:

Architectural choices depend on goals and trade-offs; this may support making architecture trade-off priorities explicit but cannot alone meet the requirement for standardizing the structure of risk assessments.

B:

Explicit extraction targets guide attention to operationally meaningful values; the stated problem requires standardizing the structure of risk assessments rather than defining bounded extraction fields.

D:

For eliciting both benefits and risks in reports, balanced section requirements make omissions more visible during validation; however, this requirement concerns standardizing the structure of risk assessments.

Question 37

A read-only documentation agent can see the full process_refund tool. What permissions change reduces accidental damage?

  1. Permit the model to answer without an obligatory tool invocation
  2. Provide the coordinator lookup access and confine refund mutations to the authorized specialist
  3. Exclude mutation tools from that agent's allowed tool set
  4. Eliminate the email-sending tool while keeping read and draft capabilities

Explanation

Correct Answer

C:

Tool access should reflect each worker's function. A documentation reviewer has no reason to possess a payment capability simply because another worker uses it.

Incorrect Answers

A:

Optional tool selection prevents unnecessary calls when the request requires no external data; that relates to avoiding forced calls on tool-free requests, not the necessary restricting tool visibility by agent role.

B:

Although partitioning capabilities prevents the orchestration role from bypassing specialist approval, the choice addresses isolating privileged refund capability and not the requested restricting tool visibility by agent role.

D:

Tool distribution should distinguish generating a proposed communication from transmitting it; the option would serve separating drafting from outbound delivery permissions, not restricting tool visibility by agent role as specified in the stem.

Question 38

A developer uses a personal experimental MCP server that must not alter teammates' repository setup. Which scope is appropriate?

  1. Depend on MCP tool discovery and declared tool interfaces
  2. Access credentials through environment or a secret manager, then rotate the leaked token
  3. Register the integration in a local or user-controlled configuration scope
  4. Set up a remote HTTP-based MCP endpoint with appropriate authentication

Explanation

Correct Answer

C:

Personal scope keeps experimental endpoints out of the shared project contract. Committing them to a repository may expose irrelevant services or configuration to the team.

Incorrect Answers

A:

MCP presents structured capabilities with names, descriptions and schemas; in this scenario, discovering MCP tools through protocol metadata is secondary to keeping personal MCP configuration out of shared files.

B:

When considering removing leaked secrets from MCP settings, configuration may be shared and reviewed; raw secrets do not belong in it; the requirement for keeping personal MCP configuration out of shared files remains unmet.

D:

Networked MCP transport supports a centrally hosted integration across clients; connecting to a shared remote MCP service and keeping personal MCP configuration out of shared files require different controls in this situation.

Question 39

An incident classifier sees both policy violations and technical outages. Which demonstration set clarifies precedence?

  1. Add a missing-field example using the schema's defined null behavior
  2. Ensure examples follow one documented schema and consistent labels
  3. Supply a few labeled input-output examples illustrating the distinction
  4. Show labeled cases where multiple classes are present and document the priority rule

Explanation

Correct Answer

D:

Mixed-category examples teach how conflicts resolve. Single-label easy cases cannot establish precedence when two categories plausibly apply.

Incorrect Answers

A:

A deliberate absent-value demonstration discourages invented totals; demonstrating missing-field semantics and teaching category precedence through examples are distinct responsibilities under these constraints.

B:

Contradictory examples undermine explicit instructions; the action supports eliminating inconsistent example output formats but falls short of teaching category precedence through examples in this case.

C:

Representative examples demonstrate how to resolve ambiguous document fields; this can serve teaching field disambiguation by example, but leaves the requirement for teaching category precedence through examples unmet.

Question 40

A resumed task appears to have applied a payment operation before the session was interrupted. What recovery step is safest?

  1. Recheck time-sensitive evidence before continuing the resumed workflow
  2. Reconcile the persisted operation ID against the payment system before retrying
  3. Sequence state-changing updates or use optimistic concurrency on the session
  4. Branch from the same validated baseline and keep branch histories separate

Explanation

Correct Answer

B:

A durable transcript does not guarantee whether the external effect committed. Idempotent reconciliation prevents double-processing during session recovery.

Incorrect Answers

A:

Because session retention preserves past context, not the current truth of external systems, the measure supports refreshing stale evidence during resume rather than the needed reconciling external actions after session interruption.

C:

Concurrent writes can produce lost updates and inconsistent tool state; supporting preventing concurrent session-state conflicts cannot replace the required reconciling external actions after session interruption.

D:

A common starting context makes alternatives comparable; this contributes to comparing alternatives from a common session baseline without meeting the need for reconciling external actions after session interruption.

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