Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 39

Question 380

A batch completes with succeeded, errored and expired records. What result-handling logic is required?

  1. Confine batch result access and enforce the organization's data retention policy
  2. Find expiration and resubmit eligible unfinished requests separately
  3. Inspect each request outcome and route failures for targeted recovery
  4. Adopt asynchronous batching and budget for deferred results

Explanation

Correct Answer

C:

An ended batch does not imply every item succeeded. Per-request result types let operators retry failed items without discarding successful work.

Incorrect Answers

A:

Bulk result files can concentrate sensitive information; its benefits for protecting sensitive bulk-processing results do not resolve the need for handling partial success within a batch.

B:

Deferred processing has time limits and can return expired outcomes; supporting recovering expired batch requests cannot replace the required handling partial success within a batch.

D:

Batch processing can reduce unit cost at the expense of immediate completion; that serves trading response latency for bulk-processing cost, whereas the case requires handling partial success within a batch.

Question 381

A support session summary must preserve case identity, unresolved human approvals, and the policy version governing pending work. Which THREE fields or checks belong in the continuity design? Choose THREE.

  1. Retain the governing refund-policy revision and revalidate rules before any pending decision
  2. Persist the verified customer case ID with the current resolution status
  3. Reduce earlier turns into a verified summary that keeps decisions and open actions
  4. Carry forward awaiting-human-approval markers, blockers and next actions explicitly

Explanation

Correct Answers

A:

This choice addresses detecting stale policy within condensed state: summaries are snapshots rather than proofs of current policy; Version-aware revalidation avoids applying a superseded refund limit to a later action.

B:

The design must include preserving stable case identity across turns: stable identifiers and status survive conversational paraphrasing; A narrative summary alone may not reliably distinguish one customer case from another.

D:

Pending workflow state is essential for safe resumption. Omitting it can cause an agent to assume authorization was granted or restart the wrong stage.

Incorrect Answers

C:

Context compaction reduces token load while retaining essential state; this may support retaining critical state while compacting history but cannot alone meet the requirement for keeping unresolved control gates in condensed context.

Question 382

The same tool call is retried after a network interruption and duplicate tickets are created. What additional interface field matters?

  1. Document typed required parameters with separate identifier semantics
  2. Make available a narrow parameterized order-status operation instead of general SQL
  3. Accept a stable idempotency key for mutation requests
  4. State explicit amount units and a required currency input

Explanation

Correct Answer

C:

A client-provided operation identifier lets the service recognize repeat submissions and return the prior outcome. Tool descriptions alone cannot make network retries safe.

Incorrect Answers

A:

A precise input schema can distinguish account identifiers from email addresses and reject unsupported shapes; the option would serve distinguishing identity fields in a tool interface, not protecting a mutating tool against replay as specified in the stem.

B:

A purpose-limited interface minimizes injection and unauthorized data access; this addresses constraining a tool to least-privilege business operations, while the question is about protecting a mutating tool against replay.

D:

Although financial tools need unambiguous units and typed fields, the choice addresses disambiguating monetary units in tool calls and not the requested protecting a mutating tool against replay.

Question 383

A multi-pass report workflow treats every detected criticism as actionable regardless of severity. What should the final pass do?

  1. Contrast cited evidence and apply an explicit resolution rubric
  2. Validate claims against sources before a separate editorial clarity pass
  3. Tag versions of prompts and rubrics and rerun a fixed regression evaluation suite
  4. Rank findings by impact and evidence strength before producing recommendations

Explanation

Correct Answer

D:

Review output should prioritize well-supported important issues. Unfiltered criticism can overwhelm owners with low-value noise and obscure real blockers.

Incorrect Answers

A:

Disagreement is information to evaluate, not a reason to average opinions blindly; reconciling contradictory reviewer findings and prioritizing findings after independent review require different controls in this situation.

B:

Fact checking and presentation optimization use different evidence and criteria; the unsatisfied requirement is prioritizing findings after independent review, not separating factual checking from writing polish.

C:

Comparable benchmarks reveal behavior changes when the system evolves; in this scenario, making multipass quality evaluations reproducible is secondary to prioritizing findings after independent review.

Question 384

A lint check passes, but a newly generated feature produces wrong results. What does the agent need to evaluate?

  1. Begin by adding the failing edge case to tests and implement the smallest correction
  2. Test failure-path tests and add explicit handling where warranted
  3. Select failing integration output to refine the patch and rerun relevant tests
  4. Verify behavioral acceptance cases rather than treating lint success as completion

Explanation

Correct Answer

D:

Linting checks style and some static issues; it does not prove runtime correctness. Scenario-based tests provide evidence about functional outputs.

Incorrect Answers

A:

Treating review feedback as a concrete test turns a discovered omission into durable coverage; that addresses incorporating edge-case feedback into tests, whereas this case calls for distinguishing static hygiene from functional success.

B:

Reliable implementation must address realistic negative inputs and dependency errors; this concerns validating negative paths after implementation, leaving the requirement for distinguishing static hygiene from functional success outstanding.

C:

Iterative development uses tool feedback as evidence; refining code based on integration-test feedback and distinguishing static hygiene from functional success are distinct responsibilities under these constraints.

Question 385

A remote MCP integration has expiring credentials, and the team's updated connection settings have not taken effect. Which TWO steps address both authentication continuity and active configuration? Choose TWO.

  1. Reload the revised MCP connection settings and verify the active remote endpoint
  2. Select capabilities from MCP tool discovery and declared tool interfaces
  3. Focus on protecting the transport with TLS and authenticate clients before sensitive tool calls
  4. Use renewable authenticated transport credentials in place of permanent embedded secrets

Explanation

Correct Answers

A:

Applying modified mcp server settings matters in this scenario: running sessions may retain existing transport state; confirming that a changed definition is actually active is more useful than assuming file edits instantly alter every connection.

D:

Remote MCP services need a secure authentication lifecycle. Hard-coded long-lived credentials are brittle and can leak in repository or deployment files.

Incorrect Answers

B:

MCP presents structured capabilities with names, descriptions and schemas; that serves discovering MCP tools through protocol metadata, whereas the case requires maintaining authenticated MCP transport.

C:

Because tool inputs and results may contain confidential operational data, the measure supports protecting an MCP network connection rather than the needed maintaining authenticated MCP transport.

Question 386

A research question contains three independent competitor comparisons. How should the coordinator divide the work?

  1. Prioritize making data normalization a prerequisite for the dependent analysis
  2. Begin by grouping closely related low-cost tasks into a meaningful worker assignment
  3. Partition read-only verification from the authorized state-changing step
  4. Assign one bounded comparison to each worker and merge on a shared schema

Explanation

Correct Answer

D:

Independent subtasks can run concurrently when they have comparable output contracts. A normalized merge stage prevents each worker from defining success differently.

Incorrect Answers

A:

The task graph must express actual data dependencies; this may support reflecting real data dependencies in the task graph but cannot alone meet the requirement for decomposing parallel independent analyses.

B:

Excessive decomposition multiplies setup and context cost; the stated problem requires decomposing parallel independent analyses rather than avoiding overfragmentation of delegated work.

C:

Distinct stages expose eligibility decisions and permit human approval before mutations; a method suited to separating validation from side-effecting tasks does not meet the need for decomposing parallel independent analyses.

Question 387

A developer uses a personal instruction to override mandatory repository security reviews. What governance distinction matters?

  1. Adopt the user's personal CLAUDE.md for cross-project preferences
  2. Corroborate the intended CLAUDE.md is discovered and loaded for that working context
  3. Maintain CLAUDE.md concise and link to task-relevant supporting references
  4. Delete secrets from instruction text and use secure environment-based credential provisioning
  5. Keep mandatory security policy in enforced controls rather than editable personal preferences

Explanation

Correct Answer

E:

Instruction hierarchy should not be the sole defense for high-risk operations. Organization-level policies and tool gates are stronger than an individual's prompt customization.

Incorrect Answers

A:

User-scoped instructions express developer preferences beyond one repository; that relates to applying personal preferences across projects, not the necessary separating organization policy from user preferences.

B:

Although rules are useful only when the assistant encounters the relevant file and scope, the choice addresses verifying instruction-file loading and not the requested separating organization policy from user preferences.

C:

High-signal standing instructions are easier to apply and less costly in context; the deciding condition is separating organization policy from user preferences, whereas this action concerns keeping persistent instructions focused.

D:

Instruction files are often shared or exposed as model context; the option would serve protecting credentials from instruction-file exposure, not separating organization policy from user preferences as specified in the stem.

Question 388

A tool-using agent makes extra unrelated queries after finishing the assigned lookup. What prompt boundary improves focus?

  1. Tie decisions and proposals to named sources, preserving unresolved disagreement
  2. Describe task scope, stopping conditions and prohibited side quests
  3. Insist on verbatim preservation for named entities while normalizing only designated fields
  4. Request source-backed claims and a labeled unknowns or uncertainty section

Explanation

Correct Answer

B:

Bounded instructions tell the agent what enough evidence looks like. Open-ended research directives can cause unnecessary retrieval and context growth.

Incorrect Answers

A:

Attribution is part of factual correctness; in this scenario, preserving attribution in generated meeting summaries is secondary to defining a stopping condition for research.

C:

When addressing preserving exact source spelling where required, field-level instructions distinguish transcription from interpretation; the present task instead calls for defining a stopping condition for research.

D:

Requiring evidence mapping and explicit uncertainty discourages treating unverified inference as fact; separating sourced conclusions from unknowns and defining a stopping condition for research require different controls in this situation.

Question 389

A coding agent needs to replace a single function inside an existing file while preserving unrelated lines. Which built-in tool best matches?

  1. Enumerate with Glob the matching migration paths and then inspect them in order
  2. Select Glob to enumerate files matching the path pattern
  3. Use Edit for a targeted modification to an existing file
  4. Choose Bash to execute the authorized test process

Explanation

Correct Answer

C:

Targeted editing changes selected text while maintaining the surrounding source. Rewriting the complete file unnecessarily raises the risk of damaging unrelated content.

Incorrect Answers

A:

Filename enumeration establishes the migration set without executing anything; gathering migration files without executing code and changing a limited region of existing code are distinct responsibilities under these constraints.

B:

Glob resolves matching file paths by name or path shape; the action supports enumerating files by filename pattern but falls short of changing a limited region of existing code in this case.

D:

Bash runs shell commands and returns their process results; that addresses executing a project test command, whereas this case calls for changing a limited region of existing code.

Question 390

An MCP tool error contains only 'failed' and the support agent guesses a fix. What response would improve diagnosis?

  1. Include a safe error category and actionable next-step detail
  2. Tag the outcome as uncertain and request reconciliation before replay
  3. Produce a non-retryable validation error with the offending field identified
  4. Emit a tool result marked isError for the failed operation

Explanation

Correct Answer

A:

Structured failure context allows the caller to distinguish bad inputs, authentication problems and outages. A generic failure string invites hallucinated remediation.

Incorrect Answers

B:

A timeout does not prove nonexecution for mutating operations; its benefits for handling an ambiguous post-request timeout do not resolve the need for making tool failure diagnostic rather than opaque.

C:

Input-shape errors persist until arguments change; supporting distinguishing validation faults from transient errors cannot replace the required making tool failure diagnostic rather than opaque.

D:

MCP distinguishes a tool execution error from a normal successful result with an error flag; this contributes to signaling execution failure in an MCP result without meeting the need for making tool failure diagnostic rather than opaque.

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