Anthropic CCAR-P (Claude Certified Architect - Professional) Practice Test 35

Question 340

Complaints circulate through chat but never become reproducible or prioritized defects. Which change should the architect recommend first?

  1. Route user feedback into a triaged, owned backlog
  2. Define SLOs and error budgets for agreed user outcomes
  3. Establish a documented escalation path for service incidents
  4. Clarify shared responsibility for the service boundary

Explanation

Correct Answer

A:

Classify feedback by severity and root-cause type, assign owners, link trace IDs, and feed validated cases into regression tests; issue lifecycle metrics show whether feedback produces controlled improvement; verify the repair using time from actionable feedback to triage, owner assignment, and verified fix.

Incorrect Answers

B:

Generic uptime claims can conceal slow or unusable answer experiences; establish measurable user-facing service objectives, reporting windows, and error-budget policies tied to accountable actions; that control is assessed by sLO compliance and prescribed actions during a simulated error-budget breach; the incident instead requires time from actionable feedback to triage, owner assignment, and verified fix.

C:

Generic 'contact support' guidance lacks accountable timing and action for critical failures; define incident severity, on-call ownership, manual fallback authority, and customer communication triggers; that control is assessed by time to appropriate escalation and fallback activation during drills; the incident instead requires time from actionable feedback to triage, owner assignment, and verified fix.

D:

A vendor SLA typically does not assign responsibility for every custom integration and decision rule; define provider, platform, application, security, and business-owner responsibilities in an explicit operating model; that control is assessed by ownership coverage for failures across model, application, data, and tool layers; the incident instead requires time from actionable feedback to triage, owner assignment, and verified fix.

Question 341

A claims assistant receives a fresh customer story each call while its authorization rules and output contract are stable. Which choice most directly satisfies the documented goal?

  1. Limit disclosure of secrets and confidential instructions
  2. Write explicit roles, scope, and prohibited operations
  3. Define an explicit abstention path for insufficient evidence
  4. Keep invariant policy instructions distinct from per-request input
  5. Use a machine-readable output contract and validator

Explanation

Correct Answer

D:

Separate stable system-level policy from user content and tool results, retaining consistent privilege boundaries across requests; boundary-specific tests confirm that changing user content cannot rewrite the application contract; evaluate the design with rate of user messages that fail to override defined task boundaries in regression tests.

Incorrect Answers

A:

Redact sensitive fields upstream, prohibit secret reproduction in instructions, and screen outbound content before release; a prohibition in the prompt alone cannot make exposed credentials safe or replace upstream data minimization; this alternative would measure secret-exposure rate under seeded credential-bearing diagnostic inputs; it does not establish rate of user messages that fail to override defined task boundaries in regression tests for the described use case.

B:

Define the assistant's permitted tasks, out-of-scope handling, and escalation behavior in stable application-level instructions, backed by independent access enforcement; role instructions improve behavior but cannot grant or restrict database permissions on their own; this alternative would measure out-of-scope refusal and handoff accuracy on a policy-boundary test set; it does not establish rate of user messages that fail to override defined task boundaries in regression tests for the described use case.

C:

Tell the assistant when to identify missing evidence, ask for clarification, or escalate, and reinforce this with evidence validation; a fallback message cannot repair a poor retrieval index; it only prevents unsupported certainty; this alternative would measure unsupported-answer rate and appropriate abstention precision on no-answer cases; it does not establish rate of user messages that fail to override defined task boundaries in regression tests for the described use case.

E:

Request a documented structured format and validate schema, allowed values, and business semantics before downstream use; human-readable examples alone do not guarantee that all runtime outputs satisfy machine contracts; this alternative would measure valid schema rate and zero unauthorized enum values reaching the consumer; it does not establish rate of user messages that fail to override defined task boundaries in regression tests for the described use case.

Question 342

Recurring clause omissions persist for months because operations never feed reviewed corrections into testing. Which action best removes the underlying failure mode?

  1. Use a queue and asynchronous workers for bursty workloads
  2. Capture reviewer corrections as labeled feedback
  3. Link each output to source and processing trace IDs
  4. Normalize incoming documents into a validated input contract

Explanation

Correct Answer

B:

Preserve authorized reviewer labels linked to case versions and turn them into regression examples with data-retention controls; retesting concrete corrections detects whether a release reintroduces known business mistakes; verify the repair using regression pass rate on previously corrected clause-omission cases.

Incorrect Answers

A:

Queueing absorbs bursts but does not improve a weak answer rubric or justify a higher-privilege tool set; decouple intake from processing through durable queues, bounded workers, and status callbacks rather than keeping user connections open; that control is assessed by job completion rate and queue age during the peak intake window; the incident instead requires regression pass rate on previously corrected clause-omission cases.

C:

Traceability explains decisions after the fact; it is not a replacement for preventing unauthorized source access; correlation IDs, source versions, and stage logs establish end-to-end lineage across extraction, retrieval, inference, and review; that control is assessed by fraction of sampled outputs reproducibly linked to input, model configuration, and reviewer disposition; the incident instead requires regression pass rate on previously corrected clause-omission cases.

D:

Input normalization addresses upstream ambiguity, not the authorization needed before writing to a finance ledger; an input-adapter layer validates document type, field types, provenance, and required metadata before the model reasons over values; that control is assessed by percentage of ingested documents meeting schema checks before any Claude call; the incident instead requires regression pass rate on previously corrected clause-omission cases.

Question 343

Users treat expected new restrictions as outages and file urgent support tickets. What is the most direct remediation?

  1. Clarify shared responsibility for the service boundary
  2. Establish a documented escalation path for service incidents
  3. Communicate release behavior changes before rollout
  4. Route user feedback into a triaged, owned backlog
  5. Define SLOs and error budgets for agreed user outcomes

Explanation

Correct Answer

C:

Define release notes, behavior-change notices, affected audiences, and support channels proportional to the deployment's impact; communication quality is reflected in predictable adoption and fewer avoidable surprises; verify the repair using affected-user awareness and support-ticket volume following staged changes.

Incorrect Answers

A:

A vendor SLA typically does not assign responsibility for every custom integration and decision rule; define provider, platform, application, security, and business-owner responsibilities in an explicit operating model; that control is assessed by ownership coverage for failures across model, application, data, and tool layers; the incident instead requires affected-user awareness and support-ticket volume following staged changes.

B:

Generic 'contact support' guidance lacks accountable timing and action for critical failures; define incident severity, on-call ownership, manual fallback authority, and customer communication triggers; that control is assessed by time to appropriate escalation and fallback activation during drills; the incident instead requires affected-user awareness and support-ticket volume following staged changes.

D:

Unstructured rating totals cannot by themselves direct a safe corrective action; classify feedback by severity and root-cause type, assign owners, link trace IDs, and feed validated cases into regression tests; that control is assessed by time from actionable feedback to triage, owner assignment, and verified fix; the incident instead requires affected-user awareness and support-ticket volume following staged changes.

E:

Generic uptime claims can conceal slow or unusable answer experiences; establish measurable user-facing service objectives, reporting windows, and error-budget policies tied to accountable actions; that control is assessed by sLO compliance and prescribed actions during a simulated error-budget breach; the incident instead requires affected-user awareness and support-ticket volume following staged changes.

Question 344

Teams ship noisy prompt changes that later regress when traffic normalizes. What is the most direct remediation?

  1. Randomize consistently at the user or account level
  2. Inspect segment-level effects rather than only global lift
  3. Release canaries with guardrail metrics and rollback conditions
  4. Predefine success criteria and sample-size assumptions
  5. Require an explicit rollback trigger for harmful regressions

Explanation

Correct Answer

D:

Define primary metrics, minimal meaningful effect, sample-size method, and stopping rules before examining results; sound statistical procedure distinguishes stable benefit from random variance; verify the repair using pre-registered experiment compliance and interval estimates at the planned sample size.

Incorrect Answers

A:

Request-by-request randomization can be invalid when outcomes depend on ongoing user interactions; assign a stable randomized unit aligned to the behavior being measured and maintain consistent treatment during the experiment; that control is assessed by balanced treatment cohorts and clean user-level exposure histories; the incident instead requires pre-registered experiment compliance and interval estimates at the planned sample size.

B:

A positive overall average does not guarantee every relevant user group benefits; compare treatment effects across predeclared critical segments and require safeguards where harm would be unacceptable; that control is assessed by resolution and safety deltas by channel and accessibility segment; the incident instead requires pre-registered experiment compliance and interval estimates at the planned sample size.

C:

A single offline score cannot show all production interactions or blast radius; deploy to a bounded cohort while observing quality, errors, and tail latency, with predefined stop and rollback thresholds; that control is assessed by canary threshold compliance and time to revert a seeded regression; the incident instead requires pre-registered experiment compliance and interval estimates at the planned sample size.

E:

Monitoring without an action threshold can document harm while allowing it to continue; establish automated and human-owned rollback triggers tied to critical error and unauthorized-action indicators; that control is assessed by time from triggered guardrail breach to verified rollback; the incident instead requires pre-registered experiment compliance and interval estimates at the planned sample size.

Question 345

A regulated engineering team requires a centrally enforced boundary on tools and operations used by coding assistants. What is the most defensible next step?

  1. Version team-wide CLAUDE.md guidance with the repository
  2. Scope MCP connections by project and user identity
  3. Use deterministic hooks to enforce required developer gates
  4. Use managed configuration to constrain dangerous permissions

Explanation

Correct Answer

D:

Use organization-managed policy and supported permission configuration so enforced restrictions cannot be casually overridden by per-user convenience settings; policy precedence tests validate that enterprise controls remain effective; evaluate the design with denial of prohibited commands after attempts to override local and project configuration.

Incorrect Answers

A:

Put reviewed repository-specific guidance in a shared CLAUDE.md file with appropriate scope and code review; a documentation file guides behavior but cannot enforce operating-system permissions or replace CI tests; this alternative would measure consistency of generated changes with repository build and test conventions across fresh sessions; it does not establish denial of prohibited commands after attempts to override local and project configuration for the described use case.

B:

Choose appropriate configuration scope and per-user or service authorization for MCP servers, avoiding shared cross-project secrets; a working connection does not prove the assistant is restricted to the user's intended project resources; this alternative would measure access denial for cross-project MCP resources under user identity tests; it does not establish denial of prohibited commands after attempts to override local and project configuration for the described use case.

C:

Configure supported hooks for deterministic checks at appropriate lifecycle points, with clear failure behavior and maintenance ownership; a written instruction alone is best-effort and does not guarantee enforcement of a critical command gate; this alternative would measure hook execution and blocked noncompliant operations under simulated bypass attempts; it does not establish denial of prohibited commands after attempts to override local and project configuration for the described use case.

Question 346

A RAG assistant will read customer-submitted documents and must not follow instructions embedded in them. What should the solution architect recommend?

  1. Adopt expert rubrics for nuanced qualitative outputs
  2. Use deterministic code checks for objective output contracts
  3. Include hostile prompt-injection and no-answer probes
  4. Stratify a golden dataset by real workload segments

Explanation

Correct Answer

C:

Add adversarial retrieved text, invalid queries, ambiguous requests, and missing-evidence cases with expected safe behavior; red-team cases make trust-boundary performance observable; evaluate the design with attack success rate and safe handling of intentionally unanswerable prompts.

Incorrect Answers

A:

Use a criterion-based expert rubric, optionally supported by calibrated model grading, for semantic accuracy and appropriateness; pure string comparison is unsuitable when many substantively correct phrasings exist; this alternative would measure expert-rated factual fidelity and calibrated grader agreement on sampled explanations; it does not establish attack success rate and safe handling of intentionally unanswerable prompts for the described use case.

B:

Apply fast programmatic validators for JSON shape, type constraints, allowed values, and deterministic business rules; subjective model grading is unnecessary and less reliable for an objectively checkable parser contract; this alternative would measure zero schema violations accepted by the downstream consumer across generated cases; it does not establish attack success rate and safe handling of intentionally unanswerable prompts for the described use case.

D:

Build a curated, versioned golden set covering volume, business risk, language, format, and difficult edge cases; a random sample alone can underrepresent important low-frequency and high-consequence events; this alternative would measure per-stratum pass rates with no uncovered critical scenario classes; it does not establish attack success rate and safe handling of intentionally unanswerable prompts for the described use case.

Question 347

The assistant responds to an out-of-scope request as though its role includes approving disciplinary action. Which change should the architect recommend first?

  1. Adopt a machine-readable output contract and validator
  2. Delimit retrieved content as untrusted evidence
  3. Write explicit roles, scope, and prohibited operations
  4. Define an explicit abstention path for insufficient evidence

Explanation

Correct Answer

C:

Define the assistant's permitted tasks, out-of-scope handling, and escalation behavior in stable application-level instructions, backed by independent access enforcement; boundary tests demonstrate task scoping rather than only fluent responses on ordinary requests; verify the repair using out-of-scope refusal and handoff accuracy on a policy-boundary test set.

Incorrect Answers

A:

Human-readable examples alone do not guarantee that all runtime outputs satisfy machine contracts; request a documented structured format and validate schema, allowed values, and business semantics before downstream use; that control is assessed by valid schema rate and zero unauthorized enum values reaching the consumer; the incident instead requires out-of-scope refusal and handoff accuracy on a policy-boundary test set.

B:

Delimiters aid instruction hierarchy but do not sanitize malicious destinations or replace output security controls; clearly separate lower-trust source material from application instructions and enforce permissions at the tool layer as well; that control is assessed by prompt-injection resistance on hostile retrieved snippets without loss of legitimate source use; the incident instead requires out-of-scope refusal and handoff accuracy on a policy-boundary test set.

D:

A fallback message cannot repair a poor retrieval index; it only prevents unsupported certainty; tell the assistant when to identify missing evidence, ask for clarification, or escalate, and reinforce this with evidence validation; that control is assessed by unsupported-answer rate and appropriate abstention precision on no-answer cases; the incident instead requires out-of-scope refusal and handoff accuracy on a policy-boundary test set.

Question 348

Claude produces plausible method signatures that no longer match the repository's actual endpoints. What is the most direct remediation?

  1. Package reusable engineering instructions as governed Skills
  2. Use test-first acceptance cases for generated changes
  3. Generate documentation from verified source and review it
  4. Review AI-generated diffs for logic and security regressions

Explanation

Correct Answer

C:

Ground documentation generation in current code, schemas, and tests, then validate examples and have owners review changes; executable examples reveal fabricated or obsolete instructions; verify the repair using documented endpoint examples executing successfully against the reviewed build.

Incorrect Answers

A:

Shared instructions improve consistency but do not replace enforced approval or automated tests; maintain task-scoped reusable Skills or playbooks with versioning, ownership, and documented prerequisites; that control is assessed by checklist completeness across independent sessions using the reviewed Skill version; the incident instead requires documented endpoint examples executing successfully against the reviewed build.

B:

Merely reading generated code for plausibility does not replace tests of business invariants; define executable tests and expected behavior before relying on AI-generated implementation, then review failures and iterate; that control is assessed by passing unit and property tests including specified rounding boundaries; the incident instead requires documented endpoint examples executing successfully against the reviewed build.

D:

Successful compilation does not establish preservation of security controls; require focused code review of the actual diff, threat-relevant tests, and approval before merging sensitive changes; that control is assessed by detection of seeded authorization regressions during peer review and CI; the incident instead requires documented endpoint examples executing successfully against the reviewed build.

Question 349

Users wait for each separate read to finish sequentially despite independent services. How should the design be corrected?

  1. Parallelize independent read-only tool calls
  2. Move slow operations into asynchronous jobs with progress states
  3. Rerank a short candidate set for ambiguous retrieval queries
  4. Route expensive reasoning to high-uncertainty cases

Explanation

Correct Answer

A:

Execute independent lookups concurrently with bounded timeouts and aggregate the results after all necessary dependencies complete; concurrency wins only if the final synthesis retains complete inputs; verify the repair using end-to-end p95 improvement with all required sources still present.

Incorrect Answers

B:

A longer synchronous timeout would still tie up request resources and expose users to unreliable waits; acknowledge an asynchronous job, persist its state, and let the client poll or receive authorized completion notifications; that control is assessed by job completion reliability and duplicate-submission rate during slow-system simulations; the incident instead requires end-to-end p95 improvement with all required sources still present.

C:

Taking every retrieved hit into context may raise token cost and distraction without raising answer quality; retrieve a bounded candidate set and rerank against the actual question to concentrate useful evidence without feeding all candidates to the model; that control is assessed by grounded-answer accuracy and retrieval-stage p95 latency at several rerank depths; the incident instead requires end-to-end p95 improvement with all required sources still present.

D:

Uniform deep reasoning does not add value to deterministic records and may increase response time; use an evaluated routing policy that sends straightforward lookups to a light path and escalates uncertain cases to more capable reasoning; that control is assessed by accuracy by difficulty segment and unit cost under p95 constraints; the incident instead requires end-to-end p95 improvement with all required sources still present.

Question 350

The orchestration graph grows while token cost doubles and ordinary cases show unchanged resolution accuracy. Which change should the architect recommend first?

  1. Bound handoffs, retries, and recursion across agents
  2. Use a coordinator to route distinct specialist tasks
  3. Delegate only specialized subtasks that add measurable value
  4. Run independent specialists in parallel with a merge contract
  5. Separate independent critique from generation

Explanation

Correct Answer

C:

Introduce a specialist only when its independent expertise, tool scope, or parallelism beats a simpler call on representative evaluations; measuring the contribution of each delegate reveals needless orchestration layers; verify the repair using marginal quality improvement per added specialist against its incremental cost and latency.

Incorrect Answers

A:

Recursive collaboration is not inherently self-terminating and cannot be made safe only by asking agents to cooperate; enforce maximum depth, handoff count, timeouts, and explicit unresolved-case escalation in the orchestrator; that control is assessed by zero unbounded cycles under adversarial unresolved tasks while retaining a useful escalation record; the incident instead requires marginal quality improvement per added specialist against its incremental cost and latency.

B:

Specialist routing adds handoffs and can be wasteful for one small, homogeneous transformation; a coordinator can dispatch scoped subagents and merge their typed outputs while restricting each specialist's context and tools; that control is assessed by completed reports with correctly scoped evidence and no cross-specialty data leakage; the incident instead requires marginal quality improvement per added specialist against its incremental cost and latency.

D:

Parallel fan-out is unsafe when tasks mutate the same state or depend on each other's evolving outputs; parallel fan-out followed by deterministic aggregation can shorten wall-clock time if each branch has its own contract and errors are handled; that control is assessed by p95 end-to-end latency and deterministic merge completeness under one-branch failures; the incident instead requires marginal quality improvement per added specialist against its incremental cost and latency.

E:

A critic agent does not repair missing source documents or guarantee that all facts are verifiable; a distinct verifier with a rubric and evidence access can challenge unsupported assertions before an external document is released; that control is assessed by verified claim support rate after independent critique on seeded unsupported drafts; the incident instead requires marginal quality improvement per added specialist against its incremental cost and latency.

Claude Certified Architect - Professional (CCAR-P) Practice Tests