Anthropic CCDV-F (Claude Certified Developer - Foundations) Practice Test 46

Question 450

A synthesized research note may be challenged and auditors must reopen the original records. What change is best aligned with the goal?

  1. Summarize intermediate reasoning and preserve outcomes
  2. Keep pointers to source documents
  3. Keep a structured conversation summary
  4. Start a new clean window after task boundary

Explanation

Correct Answer

B:

References and identifiers enable later retrieval without copying the entire source into every prompt, although pointers alone do not make an inaccessible document available at inference time.

Incorrect Answers

A:

Compaction keeps conclusions, evidence, and open questions while discarding low-value intermediate tokens; the needed capability is different—references and identifiers enable later retrieval without copying the entire source into every prompt—and overaggressive summarization can remove the details that make conclusions defensible.

C:

A summary stores salient decisions, constraints, and references in a reduced representation; however, the critical mechanism here is that references and identifiers enable later retrieval without copying the entire source into every prompt, and compression may omit nuance and needs checking for critical obligations.

D:

Separating independent sessions avoids accidental carryover of assumptions and private context; the requirement instead turns on this behavior: references and identifiers enable later retrieval without copying the entire source into every prompt; the new session must still receive any relevant persistent project policy.

Question 451

An assistant declines an unsafe request but the frontend expects a text-only success message. Which of the following would work as specified?

  1. Use parameterized queries for user values
  2. Show structured refusals or errors faithfully
  3. Preserve partial-stream status
  4. Separate citations from generated claims

Explanation

Correct Answer

B:

The client handles refusal or error states without fabricating a completed action; a crucial qualification is that hiding these states may incorrectly imply the backend executed something.

Incorrect Answers

A:

Parameters keep values separate from SQL structure and prevent injection from changing the query grammar; here, the controlling capability is that the client handles refusal or error states without fabricating a completed action, but escaping output for HTML would not prevent SQL injection.

C:

The UI marks the response incomplete and avoids committing partial structured output; by comparison, the intended solution relies on this: the client handles refusal or error states without fabricating a completed action; a few visible sentences do not establish a complete API completion.

D:

The application verifies source mapping and clearly displays provenance when available; the needed capability is different—the client handles refusal or error states without fabricating a completed action—and formatting fake-looking references beautifully does not provide evidence.

Question 452

A retrieved page says the assistant must disregard its previous directions and publish a password. Which option best fits the described workflow?

  1. Treat retrieved source text as evidence
  2. Put customer questions in user messages
  3. Treat attacker content as quoted data only
  4. Use tool schemas for action interfaces

Explanation

Correct Answer

A:

External content is data to quote or analyze only within the authorized task; its operational boundary is that a source document has no standing to instruct tool use or access policy.

Incorrect Answers

B:

The user role carries the requested task without altering the registered tool definitions; this addresses a separate concern, not the fact that external content is data to quote or analyze only within the authorized task; making every customer turn a system rule would collapse trust separation.

C:

The application preserves the page as untrusted input and excludes it from control instructions; here, the controlling capability is that external content is data to quote or analyze only within the authorized task, but a malicious instruction can still be persuasive, so execution permissions remain bounded.

D:

Tool definitions and server validation specify callable operations more reliably than prose parsing; the missing capability is more specific: external content is data to quote or analyze only within the authorized task; a natural-language command is not an adequate machine-checked action contract.

Question 453

A downstream parser sometimes rejects Claude results even though API calls return 200. Which choice should be used here?

  1. Track side-effect authorization denials
  2. Monitor retrieval coverage and citation fidelity
  3. Instrument output schema failure rate
  4. Use a validated typed response contract

Explanation

Correct Answer

C:

Tracking validation failures isolates content contract problems from transport availability; the implementation must still recognize that HTTP success is not equivalent to a valid business payload.

Incorrect Answers

A:

Auditing denied calls helps identify policy confusion or prompt injection pressure; in contrast, tracking validation failures isolates content contract problems from transport availability, while denials should remain enforced even if error metrics worsen.

B:

Measuring evidence relevance and attribution identifies grounding defects; this addresses a separate concern, not the fact that tracking validation failures isolates content contract problems from transport availability; extra context tokens cannot guarantee faithful citation of the retrieved material.

D:

Typed output constraints and independent schema validation keep data machine consumable; the missing capability is more specific: tracking validation failures isolates content contract problems from transport availability; type validity does not prove the chosen category reflects the actual evidence.

Question 454

A response contains a parseable JSON object but an enum is misspelled and a numeric field is a string. Which change would satisfy the constraint?

  1. Validate values against a schema
  2. Handle invalid responses with bounded repair
  3. Apply business rules after schema checks
  4. Treat generated SQL as untrusted text

Explanation

Correct Answer

A:

Schema validation rejects type and enumeration mismatches before downstream use; a crucial qualification is that the model being instructed to use JSON is not a runtime guarantee.

Incorrect Answers

B:

A controlled repair or regeneration step may recover a valid shape without pretending the first response was correct; by comparison, the intended solution relies on this: schema validation rejects type and enumeration mismatches before downstream use; endless retries can increase cost without resolving a faulty contract.

C:

Independent business validation enforces domain-specific ranges and account scope; the needed capability is different—schema validation rejects type and enumeration mismatches before downstream use—and pure JSON Schema cannot determine whether a payment is authorized.

D:

Query validation and read-only permissions constrain execution against approved tables and patterns, whereas the requirement calls for another behavior: schema validation rejects type and enumeration mismatches before downstream use; text generation does not certify the safety or accuracy of the SQL.

Question 455

A newly observed failure survives normal retry and recovery paths. What is the right implementation response?

  1. Correlate downstream calls with request IDs
  2. Return actionable validation errors to callers
  3. Escalate repeated unknown errors with trace context
  4. Use a circuit breaker for cascading 5xx failures

Explanation

Correct Answer

C:

Capturing sanitized request IDs and relevant version metadata makes specialist diagnosis possible; silently dropping the incident gives no route to correction.

Incorrect Answers

A:

Propagated trace identifiers let operators reconstruct a request's causal path; the missing capability is more specific: capturing sanitized request IDs and relevant version metadata makes specialist diagnosis possible; comparing timestamps manually is unreliable under concurrency.

B:

A structured 4xx response names the offending field without leaking internal implementation details; by comparison, the intended solution relies on this: capturing sanitized request IDs and relevant version metadata makes specialist diagnosis possible; a generic 500 would misleadingly suggest a server outage.

D:

A circuit breaker temporarily blocks repeated failing calls while permitting controlled recovery probes, whereas the requirement calls for another behavior: capturing sanitized request IDs and relevant version metadata makes specialist diagnosis possible; it should not be used to hide permanent programming errors.

Question 456

The team's coding agent must consistently run a particular test command and follow a documented module convention. How should this requirement be met?

  1. Use a custom slash command for a named recurring action
  2. Use CLAUDE.md for repository instructions
  3. Initialize repository guidance deliberately
  4. Place narrow guidance close to the relevant code

Explanation

Correct Answer

B:

A project instruction file provides contextual guidance when Claude Code works in that repository, although it is explanatory context rather than a security boundary enforcing tool privileges.

Incorrect Answers

A:

A named command exposes reusable prompt-driven workflow entry points inside Claude Code; the needed capability is different—a project instruction file provides contextual guidance when Claude Code works in that repository—and a command definition still depends on the session's tool permissions and context.

C:

An initialization workflow can help establish and review useful CLAUDE.md guidance for the repository; however, the critical mechanism here is that a project instruction file provides contextual guidance when Claude Code works in that repository, and generated notes must be checked before being treated as authoritative development policy.

D:

Scoped project guidance can communicate local conventions without overloading every task with unrelated instructions; the requirement instead turns on this behavior: a project instruction file provides contextual guidance when Claude Code works in that repository; conflicting guidance across scopes should be reviewed for effective precedence.

Question 457

A raw client is built against one schema and must not silently assume newer request fields are accepted. Which technical control is appropriate?

  1. Preserve future-compatible unknown fields cautiously
  2. Track the API version contract
  3. Use typed response variants where applicable
  4. Cap the number of concurrent requests

Explanation

Correct Answer

B:

Explicitly supported version settings and compatibility tests maintain a stable integration contract, although documentation examples for a different revision may not apply unchanged.

Incorrect Answers

A:

Tolerant read behavior can ignore unknown optional metadata while validating required fields; the needed capability is different—explicitly supported version settings and compatibility tests maintain a stable integration contract—and silent acceptance of unknown privileged action types would be unsafe.

C:

A discriminated union or type dispatch handles text, tool calls, and other supported blocks correctly; however, the critical mechanism here is that explicitly supported version settings and compatibility tests maintain a stable integration contract, and treating all blocks as plain strings discards structural information.

D:

A concurrency pool smooths demand and limits rate-limit errors; the requirement instead turns on this behavior: explicitly supported version settings and compatibility tests maintain a stable integration contract; throughput capacity and latency vary by model and account limits.

Question 458

A company is building its own chat UI with account-specific retrieval and backend controls. Which engineering practice is the strongest match?

  1. Use the Messages API system field for app-wide behavior
  2. Use a direct API integration for a branded customer feature
  3. Use claude.ai for end-user conversational interaction
  4. Use Claude Code for repository-scoped coding assistance

Explanation

Correct Answer

B:

The Messages API allows application code to own product behavior and its user experience; consumer chat sessions are not an app's programmable backend contract.

Incorrect Answers

A:

Server code supplies a governing system instruction separately from the user-message array; by comparison, the intended solution relies on this: the Messages API allows application code to own product behavior and its user experience; placing it inside user-supplied text exposes the rule to unintended editing and trust confusion.

C:

The hosted browser experience provides a ready-made conversational interface, whereas the requirement calls for another behavior: the Messages API allows application code to own product behavior and its user experience; it should not be assumed to inherit custom production app settings or session state.

D:

Claude Code operates in a coding environment with its own tools and repository context; however, the critical mechanism here is that the Messages API allows application code to own product behavior and its user experience, and a generic text completion endpoint does not automatically manipulate local files.

Question 459

A network disconnect interrupts generation before the final event and users still see the earlier fragments. Which technical control is appropriate?

  1. Separate citations from generated claims
  2. Require approval for write statements
  3. Escape text before inserting into HTML
  4. Preserve partial-stream status

Explanation

Correct Answer

D:

The UI marks the response incomplete and avoids committing partial structured output; in production, a few visible sentences do not establish a complete API completion.

Incorrect Answers

A:

The application verifies source mapping and clearly displays provenance when available; this addresses a separate concern, not the fact that the UI marks the response incomplete and avoids committing partial structured output; formatting fake-looking references beautifully does not provide evidence.

B:

A controlled executor enforces authorized write permissions and human approval where required; here, the controlling capability is that the UI marks the response incomplete and avoids committing partial structured output, but a syntax-valid UPDATE statement is not proof of business permission.

C:

Output encoding ensures untrusted characters display as text rather than executable DOM content; the missing capability is more specific: the UI marks the response incomplete and avoids committing partial structured output; JSON validation is not a browser security measure.

Question 460

A diagnosis requires an agent to inspect error logs and adapt its following query to what it finds. Which engineering practice is the strongest match?

  1. Ask a clarifying question when critical input is missing
  2. Observe a tool result before choosing the next action
  3. Detect and resolve conflicting worker findings
  4. Escalate after repeated inconclusive tool results

Explanation

Correct Answer

B:

Alternating action and observation permits evidence-driven choices rather than a fully predetermined plan, subject to this requirement: every loop must check for stop conditions and tool failures.

Incorrect Answers

A:

Requesting an authorized scope prevents guesses from becoming unintended side effects; in contrast, alternating action and observation permits evidence-driven choices rather than a fully predetermined plan, while an internal plan cannot legitimately invent the missing tenancy information.

C:

The coordinator compares sources, dates, and evidence before emitting a final claim; here, the controlling capability is that alternating action and observation permits evidence-driven choices rather than a fully predetermined plan, but voting by worker count is not a substitute for checking authoritative documentation.

D:

Bounded fallback returns uncertainty and a trace for human follow-up rather than fabricating certainty; the missing capability is more specific: alternating action and observation permits evidence-driven choices rather than a fully predetermined plan; extra iterations are not justified once the cost and confidence guardrails trigger.

Claude Certified Developer - Foundations (CCDV-F) Practice Tests