Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 9

Question 80

A developer has approved a two-line typo fix with an exact target location. Which workflow minimizes unnecessary planning overhead?

  1. Report the architectural trade-off and request the missing business decision
  2. Display the plan for approval before allowing write-capable execution
  3. Determine specific tests and acceptance criteria alongside the edit steps
  4. Execute the narrow approved edit directly and run the relevant check
  5. Define deployment order, checkpoints and rollback triggers before execution

Explanation

Correct Answer

D:

A bounded, well-specified low-risk change often needs no lengthy planning phase. Direct implementation is appropriate when the risk and scope are already understood.

Incorrect Answers

A:

An unresolved design choice is not safe to infer from code alone; its benefits for resolving architecture ambiguity before editing do not resolve the need for implementing a tiny well-scoped change.

B:

Sensitive migrations benefit from separating design review from mutation; supporting requiring plan approval for a migration cannot replace the required implementing a tiny well-scoped change.

C:

A useful plan describes not just changes but the evidence that will establish success; this contributes to embedding verification into the implementation plan without meeting the need for implementing a tiny well-scoped change.

E:

Because multi-service updates have dependencies and failure domains that cannot be improvised safely, the measure supports preparing release sequencing and rollback rather than the needed implementing a tiny well-scoped change.

Question 81

Thousands of batch results contain sensitive customer data. What architectural rule is important?

  1. Watch pending request limits and pace submissions by available capacity
  2. Choose asynchronous batching and budget for deferred results
  3. Restrict batch result access and enforce the organization's data retention policy
  4. Retain each request's custom_id and associated error classification

Explanation

Correct Answer

C:

Bulk result files can concentrate sensitive information. Workspace isolation, access controls and deliberate deletion are essential even when requests are independent.

Incorrect Answers

A:

For respecting throughput limits when batching, batch queues remain subject to provider limits and account constraints; however, this requirement concerns protecting sensitive bulk-processing results.

B:

Batch processing can reduce unit cost at the expense of immediate completion; that may serve trading response latency for bulk-processing cost, but it is no substitute for protecting sensitive bulk-processing results.

D:

Stable item identifiers connect failures to inputs and retry history; the stated problem requires protecting sensitive bulk-processing results rather than linking each batch error to its source record.

Question 82

A research platform needs separate legal, data and editorial analyses, parallel reads where possible, and an evidence-checking stage before synthesis. Which THREE design choices meet those requirements? Choose THREE.

  1. Parallelize the independent legal, data and editorial read-only investigations
  2. Make validated evidence a required input before report synthesis begins
  3. Give every specialist editing privileges because the coordinator can merge their changes later
  4. Start synthesis before source checking and rely on a final copyedit to remove unsupported claims
  5. Assign each specialty a bounded task, leaving cross-discipline synthesis to the coordinator

Explanation

Correct Answers

A:

The design must include parallelizing independent evidence gathering: when branches have no data dependency, parallel execution reduces waiting time while leaving the coordinator responsible for combining results and handling any branch failure.

B:

Enforcing a dependency before synthesis adds a necessary safeguard: a dependency edge enforces that the reporting worker receives validated source material; starting synthesis at the same time risks producing claims before verification is available.

E:

Focused workers can receive targeted instructions and smaller context windows. The integrator then evaluates competing outputs instead of requiring one agent to juggle incompatible objectives.

Incorrect Answers

C:

The scenario calls for independent analyses and controlled integration, not mutation by every researcher. Broad write access increases risk without providing concurrency or evidence-order guarantees.

D:

A copyedit after synthesis does not enforce a dependency on validated evidence. The workflow needs evidence review to finish before the final synthesis begins.

Question 83

A parser repeatedly exceeds the allowed number of repair attempts for one document. What policy should handle this?

  1. Check extracted references against authoritative business records
  2. Apply a bounded retry budget and move the document to an exception queue
  3. Retain the original string and flag ambiguity until locale evidence resolves it
  4. Remove duplicate entries from extracted records using source identifiers and document context

Explanation

Correct Answer

B:

Repeated retries cost money and may not improve fidelity. An exception path prevents blocking all other documents while preserving a human-review opportunity.

Incorrect Answers

A:

Referential checks establish that identifiers and relationships are real; checking referential integrity after extraction and bounding extraction repair attempts require different controls in this situation.

C:

A date parser cannot safely invent locale rules; in this scenario, handling ambiguous date interpretations is secondary to bounding extraction repair attempts.

D:

When considering removing duplicate extraction artifacts, overlapping page captures can lead to double counting; the requirement for bounding extraction repair attempts remains unmet.

Question 84

An agent receives an unclear instruction to 'remove the account' but deletion and deactivation have different consequences. What should it do?

  1. Clarify the intended action and confirm authorization before choosing either operation
  2. Describe the approval dependency and avoid claiming the refund has occurred
  3. Ask for policy-owner resolution before applying either disputed instruction
  4. Forward the case to a human approver with amount and policy evidence

Explanation

Correct Answer

A:

Destructive and reversible operations are not interchangeable. The agent must resolve the user's precise intent and obtain required approval before mutation.

Incorrect Answers

B:

Transparent status respects the user's need for clarity without misrepresenting incomplete action; this concerns communicating a pending human approval accurately, leaving the requirement for clarifying an ambiguous destructive request outstanding.

C:

Conflicting authoritative sources create genuine uncertainty; escalating conflicting governing policies and clarifying an ambiguous destructive request are distinct responsibilities under these constraints.

D:

Threshold exceptions require explicit authority; the action supports routing above-limit refunds to a human but falls short of clarifying an ambiguous destructive request in this case.

Question 85

A build engineer needs to run the repository's test command and inspect the exit status. Which built-in tool is designed for this?

  1. Search with Grep for occurrences, inspect affected files, then apply controlled Edits
  2. Choose Glob for pathname discovery before reading content
  3. Use Bash to execute the authorized test process
  4. Implement a scoped Edit with an exact matching target

Explanation

Correct Answer

C:

Bash runs shell commands and returns their process results. It should be bounded by tool permissions, particularly where commands can modify files or access secrets.

Incorrect Answers

A:

A search-first workflow reveals scope and avoids blind global replacement; this contributes to scoping a multi-file refactor before mutation without meeting the need for executing a project test command.

B:

Path globbing matches file names without requiring content search; that serves searching by path rather than file contents, whereas the case requires executing a project test command.

D:

A targeted replacement can preserve neighboring configuration content; its benefits for making a minimal configuration edit do not resolve the need for executing a project test command.

Question 86

A team tests a new support policy in a session branch without changing the approved baseline. Customer sessions must stay isolated, and forked recommendations must be checked against the live policy version. Choose THREE.

  1. Fork the approved support session into an isolated policy-testing branch
  2. Keep each customer's session and account authorization separate
  3. Check the branch's governing policy revision against the live approved version
  4. Save message order and essential tool-result linkage along with session metadata

Explanation

Correct Answers

A:

This choice addresses branching session state without overwriting the original: forking supports experimentation from a shared starting point while separating subsequent messages; appending tests to the original transcript would alter its decision history.

B:

The design must include isolating unrelated customer conversations: separate sessions prevent accidental disclosure and policy contamination between customers; carrying unrelated prior history into a new case creates privacy and correctness risks.

C:

A session branch can retain a different configuration snapshot. Explicit version reconciliation prevents merging conclusions based on incompatible policy assumptions.

Incorrect Answers

D:

For checkpointing structured agent progress, resumable state includes structured execution history, not merely the final natural-language summary; however, this requirement concerns detecting configuration drift across forks.

Question 87

A coding agent is asked to improve a function but responds with a large unrelated cleanup. What editing strategy prevents scope drift?

  1. Open the actual dependency version and local interface before rewriting calls
  2. Add the failing edge case to tests and implement the smallest correction
  3. Execute contract or compatibility tests for existing consumers
  4. Keep changes bounded to the requested behavior and a small verified diff

Explanation

Correct Answer

D:

Narrow modifications are easier to review, test and revert. Unrelated refactoring expands regression surface without helping the acceptance criteria.

Incorrect Answers

A:

Version drift can make a plausible call invalid; that relates to aligning generated code to installed dependency APIs, not the necessary limiting a patch to the requested behavior.

B:

Although treating review feedback as a concrete test turns a discovered omission into durable coverage, the choice addresses incorporating edge-case feedback into tests and not the requested limiting a patch to the requested behavior.

C:

Code that compiles can still break callers relying on prior inputs or outputs; the deciding condition is limiting a patch to the requested behavior, whereas this action concerns testing API compatibility after code generation.

Question 88

An agent scheduler launches too many workers at once and lets them run without an exit criterion. Which TWO changes address both capacity and completion control? Choose TWO.

  1. Give each scheduled worker a bounded job and explicit stop or escalation condition
  2. Focus on completing mandatory high-value deliverables before optional expansions
  3. Allocate one bounded comparison to each worker and merge on a shared schema
  4. Queue additional worker tasks behind a configured concurrency ceiling

Explanation

Correct Answers

A:

Defining bounded task exits adds a necessary safeguard: bounded tasks define what evidence and action terminate work; an open-ended instruction can lead to unproductive loops and unclear human handoff.

D:

Capacity-aware scheduling controls rate limits and tool contention. Unlimited parallelism can increase throttling and retries rather than speed up completion.

Incorrect Answers

B:

When considering prioritizing the critical reporting path, a critical-path plan ensures that the required outcome exists even if time runs short; the requirement for scheduling within available execution capacity remains unmet.

C:

When addressing decomposing parallel independent analyses, independent subtasks can run concurrently when they have comparable output contracts; the present task instead calls for scheduling within available execution capacity.

Question 89

A cancellation step fails after inventory was reserved successfully. What should the flow plan provide?

  1. Introduce an executable policy gate before the final approval or refund call
  2. Start by making eligibility verification a required predecessor in the workflow graph
  3. Model compensation or recovery actions for partial multi-step failure
  4. Send to identity clarification instead of advancing to account-specific tools

Explanation

Correct Answer

C:

External systems rarely offer one transaction across all tools. Compensation paths restore consistency or flag manual cleanup when an earlier side effect cannot simply be rolled back.

Incorrect Answers

A:

A deterministic guard can compare amounts with configured limits; enforcing financial limits in a workflow and recovering from a partially committed workflow are distinct responsibilities under these constraints.

B:

An explicit dependency ensures the mutation cannot begin before policy checks succeed; the action supports enforcing refund prerequisites before mutation but falls short of recovering from a partially committed workflow in this case.

D:

Identity-dependent steps require verified identifiers; that addresses blocking state transitions on missing identity, whereas this case calls for recovering from a partially committed workflow.

Question 90

A malformed JSON payload prevents an MCP tool handler from parsing the request. What should the server provide?

  1. Return a structured invalid-arguments failure with an expected-shape hint
  2. Label the business constraint as nontransient and require a different next step
  3. Mark partial completion explicitly and identify which data remains unavailable
  4. Add a safe error category and actionable next-step detail

Explanation

Correct Answer

A:

Errors about parsing and schema constraints should identify how the caller can repair the input. Pretending the operation succeeded would poison the workflow with false data.

Incorrect Answers

B:

Funds or policy failures generally need user or system-state changes, not immediate identical retries; its benefits for avoiding futile retries on business-rule rejection do not resolve the need for explaining invalid tool arguments.

C:

A partial response must not masquerade as complete evidence; supporting communicating incomplete tool output cannot replace the required explaining invalid tool arguments.

D:

Structured failure context allows the caller to distinguish bad inputs, authentication problems and outages; this contributes to making tool failure diagnostic rather than opaque without meeting the need for explaining invalid tool arguments.

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