Anthropic CCDV-F (Claude Certified Developer - Foundations) Practice Test 29

Question 280

A large research task has multiple deliverables and progress is becoming hard to verify. Which of these alternatives serves the stated goal?

  1. Segment work into explicit milestones
  2. Use an external source refresh for changing facts
  3. Keep unresolved questions explicit
  4. Preserve immutable business facts in structured state

Explanation

Correct Answer

A:

Milestones with acceptance criteria limit aimless elaboration and provide reliable checkpoints; the implementation must still recognize that a plan without tested completion criteria will still drift.

Incorrect Answers

B:

Refetching authoritative information avoids treating outdated context as eternally current; the missing capability is more specific: milestones with acceptance criteria limit aimless elaboration and provide reliable checkpoints; the model should not invent a change simply because its history is old.

C:

Recording open issues prevents compaction from converting uncertainty into an invented guarantee; by comparison, the intended solution relies on this: milestones with acceptance criteria limit aimless elaboration and provide reliable checkpoints; suppressing gaps for a cleaner paragraph undermines factual reliability.

D:

The application records authoritative facts outside model-generated summaries and supplies them when needed; the needed capability is different—milestones with acceptance criteria limit aimless elaboration and provide reliable checkpoints—and conversation recollection can drift or omit earlier commitments.

Question 281

An assistant chooses between similar inventory tools and submits nested item quantities to an order endpoint. The team wants discoverable contracts and machine-rejected malformed arguments. Which TWO improvements target those weaknesses? Choose TWO.

  1. Mark truly required arguments in the schema
  2. Use a client-executed custom tool
  3. Constrain nested object shapes
  4. Provide precise tool descriptions and argument types
  5. Represent allowed values as enums where appropriate

Explanation

Correct Answers

C:

Structured nested validation rejects incomplete or malformed records consistently; in production, accepting arbitrary object shapes can produce unpredictable writes.

D:

Clear names, schema fields, and descriptions help the model choose and populate the intended operation; tool schemas still need server-side validation of all inputs.

Incorrect Answers

A:

The schema identifies mandatory fields so calls missing them can be rejected before dispatch; here, the controlling capability is that structured nested validation rejects incomplete or malformed records consistently, but optional property defaults should not hide missing business-critical identifiers.

B:

The API returns a tool_use request and the application performs the authorized operation itself; the missing capability is more specific: structured nested validation rejects incomplete or malformed records consistently; tool declarations do not grant the vendor access to private infrastructure.

E:

Constrained value sets prevent unsupported strings from reaching backend business logic, whereas the requirement calls for another behavior: structured nested validation rejects incomplete or malformed records consistently; valid enum membership alone cannot authorize an illegal state transition.

Question 282

Before any implementation, product and support teams must agree on what users are told and how requests are recorded during a complete model-provider outage. Which choice should be used here?

  1. Specify source attribution for factual answers
  2. Define the session continuity model
  3. Define fallback behavior under model outage
  4. Define the authority of each tool action

Explanation

Correct Answer

C:

Availability requirements identify queueing, safe degraded responses, and recovery after service restoration, although blindly waiting for an API call does not guarantee a usable service.

Incorrect Answers

A:

Product requirements describe links or identifiers for supporting evidence and how claims map to sources; by comparison, the intended solution relies on this: availability requirements identify queueing, safe degraded responses, and recovery after service restoration; requiring a confident tone does not establish provenance.

B:

Session rules identify which prior context persists and when it must be invalidated; the needed capability is different—availability requirements identify queueing, safe degraded responses, and recovery after service restoration—and accidentally reusing a cross-account history can expose sensitive context.

D:

Access requirements tie each permitted side effect to identity, scope, and approval evidence; however, the critical mechanism here is that availability requirements identify queueing, safe degraded responses, and recovery after service restoration, and a model's confidence in a request is not proof of caller authority.

Question 283

Offline report generation can be delayed without losing incoming customer requests. Which mechanism should the team prioritize?

  1. Return a verified FAQ when inference is unavailable
  2. Reject unsupported operations explicitly
  3. Queue nonurgent tasks during provider outages
  4. Use a background job worker for long runs

Explanation

Correct Answer

C:

Durable storage retains jobs for resumed processing once inference is available; an in-memory list in a failed worker may lose pending work.

Incorrect Answers

A:

A controlled cache of approved responses preserves basic service where exact matches exist; the missing capability is more specific: durable storage retains jobs for resumed processing once inference is available; the system must avoid pretending a cached answer is newly generated or universally applicable.

B:

A clear capability boundary returns an accurate unsupported-action response; by comparison, the intended solution relies on this: durable storage retains jobs for resumed processing once inference is available; inventing a successful action would violate user trust and corrupt workflow state.

D:

Worker queues provide durable processing and status reporting outside the initial HTTP connection, whereas the requirement calls for another behavior: durable storage retains jobs for resumed processing once inference is available; synchronous endpoints can time out under long processing loads.

Question 284

A tool receives a temporary 503 from an upstream search service during an agent run. Which choice should be used here?

  1. Check tool arguments against a schema
  2. Return a tool_result with the matching tool_use_id
  3. Retry transient failures with backoff
  4. Avoid returning unnecessary raw records

Explanation

Correct Answer

C:

A bounded retry policy handles intermittent service errors without indefinitely hammering the dependency—with this limitation: repeating a rejected unauthorized request cannot repair missing permission.

Incorrect Answers

A:

Typed validation rejects malformed or out-of-policy arguments before backend execution, whereas the requirement calls for another behavior: a bounded retry policy handles intermittent service errors without indefinitely hammering the dependency; a model-generated JSON object can be syntactically valid while still violating business rules.

B:

The caller adds a user-side tool_result block tied to the earlier tool use identifier; however, the critical mechanism here is that a bounded retry policy handles intermittent service errors without indefinitely hammering the dependency, and an unrelated message cannot reliably bind a result to the invocation it completes.

D:

The tool should return task-relevant, minimized structured data instead of dumping unneeded sensitive details; this cannot substitute for the behavior that a bounded retry policy handles intermittent service errors without indefinitely hammering the dependency, because excessive outputs increase context costs and exposure while burying the useful evidence.

Question 285

A Messages response may include multiple block types with distinct fields and meanings. Which approach best meets this requirement?

  1. Negotiate supported content types properly
  2. Track the API version contract
  3. Use typed response variants where applicable
  4. Associate deltas with their content block indices

Explanation

Correct Answer

C:

A discriminated union or type dispatch handles text, tool calls, and other supported blocks correctly; treating all blocks as plain strings discards structural information.

Incorrect Answers

A:

Content-type checks help the parser distinguish valid API data from an unexpected gateway body; the missing capability is more specific: a discriminated union or type dispatch handles text, tool calls, and other supported blocks correctly; blind JSON parsing obscures transport failures.

B:

Explicitly supported version settings and compatibility tests maintain a stable integration contract; by comparison, the intended solution relies on this: a discriminated union or type dispatch handles text, tool calls, and other supported blocks correctly; documentation examples for a different revision may not apply unchanged.

D:

Block-index tracking assembles each text or tool input stream correctly, whereas the requirement calls for another behavior: a discriminated union or type dispatch handles text, tool calls, and other supported blocks correctly; appending every delta to a single string can mix text with tool parameters.

Question 286

A service needs a stable role and behavioral policy separate from customer utterances. Which capability is necessary for this outcome?

  1. Set an explicit max_tokens budget
  2. Choose a supported model identifier
  3. Use the top-level system instruction field
  4. Parse typed content blocks rather than raw concatenation

Explanation

Correct Answer

C:

The Messages API represents governing instructions at request level rather than as a fabricated system role in messages; the important constraint is that treating a system instruction as untrusted user content can blur instruction precedence.

Incorrect Answers

A:

The request parameter limits output generation independently from the size of the prompt; the needed capability is different—the Messages API represents governing instructions at request level rather than as a fabricated system role in messages—and a larger max_tokens value does not expand the model's input context window.

B:

The model field identifies the deployed model; production should verify current IDs and compatibility, whereas the requirement calls for another behavior: the Messages API represents governing instructions at request level rather than as a fabricated system role in messages; guessing an informal nickname can yield an invalid-model error or unexpected alias changes.

D:

Typed block handling preserves whether an element is displayable text or a tool request; the requirement instead turns on this behavior: the Messages API represents governing instructions at request level rather than as a fabricated system role in messages; blindly concatenating heterogeneous blocks can lose tool metadata and semantics.

Question 287

An investigation accumulated repeated search hits that no longer inform the current decision. What is the correct engineering decision?

  1. Keep a structured conversation summary
  2. Remove stale tool outputs
  3. Retrieve relevant evidence per user question
  4. Use targeted tool requests instead of bulk dumps

Explanation

Correct Answer

B:

Pruning outdated results frees context for more relevant evidence and current questions; its operational boundary is that removing the sole authoritative citation would weaken the final answer.

Incorrect Answers

A:

A summary stores salient decisions, constraints, and references in a reduced representation; this cannot substitute for the behavior that pruning outdated results frees context for more relevant evidence and current questions, because compression may omit nuance and needs checking for critical obligations.

C:

Retrieval selects bounded passages pertinent to the current question and supplies their source context; this addresses a separate concern, not the fact that pruning outdated results frees context for more relevant evidence and current questions; selecting on superficial keywords alone may miss nuanced exceptions.

D:

Narrow queries reduce context consumption and sensitive data exposure simultaneously; here, the controlling capability is that pruning outdated results frees context for more relevant evidence and current questions, but returned results still need authorization before the query executes.

Question 288

A background investigative agent must never run beyond a documented service-level deadline. Which solution should the engineer adopt?

  1. Apply a per-run timeout
  2. Apply a sandbox to untrusted generated code
  3. Set a maximum tool iteration count
  4. Require explicit approval for irreversible actions

Explanation

Correct Answer

A:

The host checks elapsed time and cancels or stops further tool iterations at the limit—with this limitation: increasing token capacity does not impose a real execution deadline.

Incorrect Answers

B:

Isolation constrains filesystem, processes, secrets, and network access independently of model reasoning; this cannot substitute for the behavior that the host checks elapsed time and cancels or stops further tool iterations at the limit, because a prompt claiming code is safe cannot replace runtime containment.

C:

A harness enforces a fixed ceiling on action cycles and reports incomplete execution safely; in contrast, the host checks elapsed time and cancels or stops further tool iterations at the limit, while model self-assessment is too probabilistic to serve as the only loop brake.

D:

The side-effect layer blocks execution until an authorization token or confirmation arrives; this addresses a separate concern, not the fact that the host checks elapsed time and cancels or stops further tool iterations at the limit; labeling a step dangerous in the prompt does not implement authorization.

Question 289

A quotation assistant must calculate sales tax using a fixed jurisdictional formula with auditable rounding. Which technical control is appropriate?

  1. Use a human approval checkpoint
  2. Use deterministic code for a known calculation
  3. Assign tool side effects to an execution worker
  4. Use a single supervising agent

Explanation

Correct Answer

B:

Ordinary code computes the numerical result repeatably rather than trusting text generation for arithmetic, although a language model may describe the calculation but should not be its financial source of truth.

Incorrect Answers

A:

Execution pauses before the consequential action and resumes only after an authorized reviewer agrees; the needed capability is different—ordinary code computes the numerical result repeatably rather than trusting text generation for arithmetic—and a natural-language promise to ask permission is weaker than an enforced approval gate.

C:

A tightly permissioned executor maps authorized operations into deterministic tool calls; however, the critical mechanism here is that ordinary code computes the numerical result repeatably rather than trusting text generation for arithmetic, and broad access in the planning component defeats the intended separation of duties.

D:

A supervisor holds the overall objective and coordinates narrower workers through explicit handoffs; the requirement instead turns on this behavior: ordinary code computes the numerical result repeatably rather than trusting text generation for arithmetic; its instructions and budget must prevent cycles and duplicate work among workers.

Question 290

Company policy requires all custom business tools to run inside an approved network segment. What is the appropriate next design choice?

  1. Use a local function dispatcher for private integrations
  2. Start a fresh isolated context for a subtask
  3. Run the harness in a controlled private environment
  4. Apply a sandbox to untrusted generated code

Explanation

Correct Answer

C:

A self-operated worker provides network and runtime control for executing the permitted functions; in production, the operator must maintain patching, scaling, observability, and worker isolation.

Incorrect Answers

A:

Application code executes a narrowly defined local tool and returns selected results to Claude; here, the controlling capability is that a self-operated worker provides network and runtime control for executing the permitted functions, but the API itself does not have arbitrary access to a customer's private internal network.

B:

A separate worker receives only the necessary files and reports a minimal result; the missing capability is more specific: a self-operated worker provides network and runtime control for executing the permitted functions; isolation only works when tool rights and returned text follow the same boundary.

D:

Isolation constrains filesystem, processes, secrets, and network access independently of model reasoning; the needed capability is different—a self-operated worker provides network and runtime control for executing the permitted functions—and a prompt claiming code is safe cannot replace runtime containment.

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