Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 25

Question 240

Claude Code is exploring a large monorepo for a maintenance task. The engineer knows the exact configuration file and needs to inspect one section. What should be implemented at the appropriate boundary?

  1. Use an appropriate file-edit operation with a specific expected match
  2. Use a file glob or filename discovery operation
  3. Use a targeted content search with a bounded file scope
  4. Run the build or test using a controlled shell or task tool
  5. Use a file-read operation on the known path and relevant range

Explanation

Correct Answer

E:

Direct reading preserves surrounding context when a file is already identified and avoids broad repository scans. Here it serves the explicit need to read the identified file directly without adding a separate unverified inference.

Incorrect Answers

A:

When an authorized edit must replace a precisely identified line in a tracked file, this would be useful because a focused edit is easier to review and verify than regenerating an entire file from model memory; this item instead calls for a way to read the identified file directly.

B:

Glob-style search identifies paths by name pattern without reading or executing each file in the repository; the relevant precondition would be an engineer knows a filename pattern but not its directory. The question instead describes the engineer knows the exact configuration file and needs to inspect one section.

C:

Text search finds matching symbols inside files; filename discovery alone cannot identify call sites; the alternative would make sense if an engineer needs every reference to a deprecated function in source text. It does not directly fix the observed condition: the engineer knows the exact configuration file and needs to inspect one section.

D:

This measure fits a different trigger: a scripted maintenance operation must compile the project and observe test output. In that setting, Only an actual process invocation establishes compiler and test behavior; inspection tools cannot substitute for execution; it does not resolve the present need to read the identified file directly.

Question 241

An organization publishes MCP tools for searching internal technical documentation. A file-read tool accepts any path string and can access outside the workspace. Which option best meets the operational need?

  1. Validate and constrain readable paths in the tool implementation
  2. Specify its corpus, query shape, and result guarantees in the description
  3. Describe deterministic lookup versus semantic search use cases explicitly
  4. Document that search outputs are excerpts with source references and may need follow-up reads

Explanation

Correct Answer

A:

Descriptions help selection but do not provide security; a server-side allowlist must protect unauthorized files. The relevant requirement is to enforce the workspace boundary at execution, and this measure operates at that decision point.

Incorrect Answers

B:

Model selection depends on clear capability metadata; an opaque name leaves the assistant guessing which source will answer the user's question; the relevant precondition would be a tool named query has no indication of which index it searches. The question instead describes a file-read tool accepts any path string and can access outside the workspace.

C:

When two tools share a capability, but one is optimized for exact identifier lookups, this would be useful because a direct identifier lookup gives precise retrieval when a key is known; embedding search is better for conceptual matching, not exact ID verification; this item instead calls for a way to enforce the workspace boundary at execution.

D:

An explicit limitation discourages Claude from treating a truncated hit as authoritative coverage of a whole document; the alternative would make sense if a semantic-search tool returns ranked snippets that may not be complete documents. It does not directly fix the observed condition: a file-read tool accepts any path string and can access outside the workspace.

Question 242

A customer-support assistant must decide when to request human assistance. A refund would exceed the support agent's approved limit. What is the most appropriate design choice?

  1. Apply the verified rule within delegated permission and record the decision
  2. Create a human-approval request with the verified amount and case evidence
  3. Mark the policy gap and hand off to an authorized human
  4. Trigger the organization's urgent human handoff process

Explanation

Correct Answer

B:

Authority limits must be observed at execution; the model cannot increase its own financial permissions. This matches the workflow goal to route an over-limit action to proper approval rather than introducing an unrelated action.

Incorrect Answers

A:

When a support specialist finds the policy clear and customer eligibility already verified, this would be useful because human review should be focused on risk or uncertainty; routine well-defined cases can proceed under controlled authorization; this item instead calls for a way to route an over-limit action to proper approval.

C:

The model should not create a new rule for an unaddressed case; a documented gap requires accountable policy interpretation; the relevant precondition would be a request falls outside every published returns-policy category. The question instead describes a refund would exceed the support agent's approved limit.

D:

This measure fits a different trigger: a customer reports urgent harm that the assistant cannot safely assess. In that setting, High-impact, ambiguous reports warrant a trained responder and documented procedure rather than an autonomous improvised resolution; it does not resolve the present need to route an over-limit action to proper approval.

Question 243

An application extracts structured fields from invoices with Claude. Some invoice records omit an optional tax identifier. Which response is most reliable?

  1. Have Claude supply fields through a tool interface with a defined JSON Schema
  2. Define an array of line-item objects with required amount and description fields
  3. Declare amount as a number and reject nonnumeric values
  4. Model tax ID as optional or nullable according to downstream semantics

Explanation

Correct Answer

D:

A schema that requires nonexistent information encourages fabricated values; absence must be representable explicitly.

Incorrect Answers

A:

Tool-argument schemas provide a clearer machine-readable contract than merely asking for 'JSON only' in prose, though validation remains necessary; the relevant precondition would be free-form JSON text frequently contains trailing commentary that breaks parsing. The question instead describes some invoice records omit an optional tax identifier.

B:

This measure fits a different trigger: the system stores several line items for one invoice. In that setting, Nested arrays preserve line-item boundaries and permit validation of each element without relying on free-text delimiters; it does not resolve the present need to represent optionality accurately in the schema.

C:

When a paid amount is emitted as a phrase such as 'about fifty', this would be useful because type constraints prevent textual approximations from being silently interpreted as ledger-ready amounts; this item instead calls for a way to represent optionality accurately in the schema.

Question 244

A document classifier uses Claude to tag contracts with legal categories. Reviewers identify recurring false positives from harmless boilerplate. What should the team do to resolve this particular gap?

  1. Choose examples from different layouts and writing styles while keeping the rubric stable
  2. Show benign text that resembles the target but does not satisfy its criteria
  3. Select diverse boundary cases based on observed error patterns
  4. Include labeled examples that explain how to break ties

Explanation

Correct Answer

B:

Negative cases calibrate the decision boundary more directly than adding more obvious positive examples. The relevant requirement is to add targeted negative demonstrations, and this measure operates at that decision point.

Incorrect Answers

A:

Varied examples prevent the model from associating a label with incidental formatting instead of the clause meaning; the alternative would make sense if the training examples use only one document style, but production inputs vary. It does not directly fix the observed condition: reviewers identify recurring false positives from harmless boilerplate.

C:

When the team has hundreds of redundant examples in a single prompt, this would be useful because a smaller curated set often gives clearer task guidance and avoids burying the decision rule under duplicated demonstrations; this item instead calls for a way to add targeted negative demonstrations.

D:

Examples should show which category wins at overlapping boundaries rather than leaving Claude to infer an unstated hierarchy; the relevant precondition would be a particular clause belongs to two related categories but reviewers expect one primary label. The question instead describes reviewers identify recurring false positives from harmless boilerplate.

Question 245

A document-processing agent invokes external services while extracting invoice fields. A payment API can create a real charge, but model responses may be replayed. Which control is most directly relevant?

  1. Validate tool output against the expected result contract before reasoning
  2. Apply a retry cap and bounded backoff outside the model
  3. Gate the payment tool behind authorization and confirmation before execution
  4. Validate tool-call inputs against the declared argument schema
  5. Track pending work in durable application state until the actual result arrives

Explanation

Correct Answer

C:

Prompt wording alone is not a reliable transaction guard; the execution layer must verify that the action is approved.

Incorrect Answers

A:

An empty result may reflect OCR failure, not an intentionally blank invoice field; validation stops it being promoted to fact; the relevant precondition would be the OCR service sometimes returns an empty string rather than recognized text. The question instead describes a payment API can create a real charge, but model responses may be replayed.

B:

An application-owned retry policy prevents infinite loops and makes downstream reliability predictable under outages; the alternative would make sense if an extraction workflow repeats one expensive tool call after a transient error. It does not directly fix the observed condition: a payment API can create a real charge, but model responses may be replayed.

D:

When a tool schema requires a numeric invoice total but Claude supplies a string, this would be useful because schema checking at the host boundary prevents type errors and unsafe side effects from malformed model-generated arguments; this item instead calls for a way to make mutation execution conditional on approval.

E:

This measure fits a different trigger: tool execution takes several minutes, but the model has already emitted the request. In that setting, A long-running call must not be represented as success before completion, especially where later steps depend on its output; it does not resolve the present need to make mutation execution conditional on approval.

Question 246

An operations team monitors an agent-based customer workflow in production. The team wants to compare latency of serial and parallel research designs. How should the workflow be adjusted?

  1. Record versions, response digests, and safe replay artifacts for important external results
  2. Verify the expected task-graph completion states before marking the request successful
  3. Correlate model turns, tool requests, task IDs, and tool outcomes in structured traces
  4. Instrument fan-out, join, and worker durations separately

Explanation

Correct Answer

D:

Per-stage timing shows whether concurrency removes a real bottleneck or merely moves latency to a saturated shared service.

Incorrect Answers

A:

When a production incident cannot be reproduced because tool outputs changed since execution, this would be useful because a stable record ties decisions to the data actually observed while avoiding the assumption that a live tool will return the same content later; this item instead calls for a way to measure waiting and execution time by task edge.

B:

A polished model response is not proof that required delegated checks completed; orchestration state supplies that guarantee; the alternative would make sense if a final reply is fluent although a required specialist never ran. It does not directly fix the observed condition: the team wants to compare latency of serial and parallel research designs.

C:

End-to-end correlation reveals the sequence leading to a failure without relying on a final narrative that may omit intermediate decisions; the relevant precondition would be an agent makes repeated tool calls but the trace does not show why each was chosen. The question instead describes the team wants to compare latency of serial and parallel research designs.

Question 247

A software team builds a two-pass architecture for Claude-generated code changes. A reviewer flags missing tests but the generator reruns without using the feedback. Which implementation decision is the best fit?

  1. Provide the concrete patch, affected code, and measurable review checklist
  2. Require maintainer authorization after automated review and testing
  3. Compare both claims to the public API contract and escalate unresolved risk
  4. Return specific failing criteria and request a bounded patch update

Explanation

Correct Answer

D:

A corrective loop improves only if the generator receives actionable evidence rather than a generic negative verdict.

Incorrect Answers

A:

A reviewer needs artifact evidence and evaluation standards to assess behavior, not a vague description from the writer; the alternative would make sense if the reviewer says 'looks good' without checking the actual diff. It does not directly fix the observed condition: a reviewer flags missing tests but the generator reruns without using the feedback.

B:

When the final release decision must be owned by a named maintainer, this would be useful because automation may gather evidence, but the stated governance rule makes a person accountable for promotion to production; this item instead calls for a way to feed targeted defects back into the next revision.

C:

This measure fits a different trigger: two reviewers produce conflicting assessments of a breaking API change. In that setting, Conflicting model opinions should be resolved using independent criteria and executable evidence, not majority fluency; it does not resolve the present need to feed targeted defects back into the next revision.

Question 248

A Claude Code repository is organized into multiple services with different local rules. An imported policy file is no longer present after a directory move. What should the architect specify in the design?

  1. Keep common conventions at root and isolate service-specific guidance
  2. Update the path or import so Claude Code can load the maintained guidance
  3. Delete or archive obsolete operational details from shared instructions
  4. Use a supported CLAUDE.md import or concise reference to the maintained document

Explanation

Correct Answer

B:

A prompt reminder cannot replace a broken import; resolving the file reference restores the policy's discoverability.

Incorrect Answers

A:

Layered instructions reduce irrelevant context while preserving shared policies that apply across the repository; the relevant precondition would be the root instructions list every service's commands and overwhelm simple tasks. The question instead describes an imported policy file is no longer present after a directory move.

C:

This measure fits a different trigger: the team's root file contains a temporary incident note from last month. In that setting, Outdated incident context can mislead future tasks; persistent memory should contain durable conventions and verified references; it does not resolve the present need to repair the actual instruction source.

D:

Importing maintained guidance avoids parallel copies that drift when the release process changes; the alternative would make sense if a release policy document already exists and should not be copied into multiple instruction files. It does not directly fix the observed condition: an imported policy file is no longer present after a directory move.

Question 249

A multi-agent research product must collect, compare, and publish findings. A narrative must cite calculations derived from a completed dataset. Which change most directly addresses this condition?

  1. Sequence data collection, computation, and writing according to their dependencies
  2. Ask the coordinator to compare provenance and escalate unresolved conflicts
  3. Give the coordinator routing and synthesis responsibilities, not a second full search workload
  4. Use a coordinator to delegate scoped jobs to search, analysis, and writing workers

Explanation

Correct Answer

A:

Publishing before numbers are validated creates unsupported claims; a dependency graph makes prerequisite completion explicit.

Incorrect Answers

B:

When two specialists disagree over a key statistic from different documents, this would be useful because the coordinator should reconcile using source quality and dates, not silently select whichever worker sounds more confident; this item instead calls for a way to ensure analysis precedes final drafting.

C:

This measure fits a different trigger: a coordinator routinely duplicates its specialists' detailed searches. In that setting, Overlapping roles waste tokens and introduce conflicting versions of evidence; ownership should be explicit in the orchestration contract; it does not resolve the present need to ensure analysis precedes final drafting.

D:

A central orchestrator can assign responsibilities and synthesize outputs while specialists work with only relevant tools and context; the relevant precondition would be search, quantitative analysis, and writing each require different expertise and inputs. The question instead describes a narrative must cite calculations derived from a completed dataset.

Question 250

A document-processing agent decides whether to retry failed MCP calls. An XML converter receives a password-protected file it cannot read. What is the most defensible next step?

  1. Treat missing approval as a policy precondition requiring escalation
  2. Use an operation ID to query authoritative completion status
  3. Return a field-level constraint violation for the caller to repair
  4. Use retry-aware error metadata with capped exponential backoff
  5. Return a password-protection status rather than a fabricated empty conversion

Explanation

Correct Answer

E:

A clear input limitation enables the user to provide an accessible document through approved channels; repeated conversion is ineffective.

Incorrect Answers

A:

Approval cannot be manufactured by model retries; the workflow must obtain an authorized decision outside the model; the alternative would make sense if a tool reports that the user's approval is missing for a payout. It does not directly fix the observed condition: an XML converter receives a password-protected file it cannot read.

B:

This measure fits a different trigger: a network timeout leaves the outcome of a state-changing operation uncertain. In that setting, Retrying a financial mutation blindly can duplicate a transaction; checking durable state is safer than inferring failure from a timeout; it does not resolve the present need to request a permitted input remedy.

C:

When a validator rejects a confidence value outside its supported range, this would be useful because a type or range error requires argument correction, not infrastructure backoff or an alternate source search; this item instead calls for a way to request a permitted input remedy.

D:

Temporary saturation may clear; paced retries can recover without causing a synchronized load spike; the relevant precondition would be a service responds with a temporary capacity error and a recommended wait. The question instead describes an XML converter receives a password-protected file it cannot read.

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