Anthropic CCDV-F (Claude Certified Developer - Foundations) Practice Test 53

Question 520

A client needs a reliable template for summarizing a policy document published by a reusable MCP server. Which approach targets the deciding condition?

  1. Combine a guarded custom tool with a human approval step
  2. Combine a deterministic hook with a coding agent
  3. Combine MCP resources with a summarization prompt
  4. Use an MCP server for a reusable integration

Explanation

Correct Answer

C:

Resources supply source data while a prompt governs how it should be analyzed, although neither component independently guarantees the summary is factually faithful.

Incorrect Answers

A:

The tool executes a constrained operation after an enforced business approval gate; by comparison, the intended solution relies on this: resources supply source data while a prompt governs how it should be analyzed; model-generated consent text is not an auditable approval event.

B:

A hook enforces the external check while the model handles the flexible development task; the needed capability is different—resources supply source data while a prompt governs how it should be analyzed—and without correct hook wiring and tool permissions, guidance alone is insufficient.

D:

A protocol server exposes shared tools and resources with consistent discovery semantics; however, the critical mechanism here is that resources supply source data while a prompt governs how it should be analyzed, and it adds infrastructure and identity concerns compared with a single local function.

Question 521

A permitted operation cannot locate the referenced object and the client needs to recover intelligently. Which of these alternatives serves the stated goal?

  1. Validate operations in the server handler
  2. Advertise tool metadata clearly
  3. Use Streamable HTTP for remote MCP access
  4. Return structured errors from tools

Explanation

Correct Answer

D:

Explicit error signaling distinguishes tool failure from an empty successful result, subject to this requirement: swallowing errors can cause a model to hallucinate completion.

Incorrect Answers

A:

Server handlers enforce identity, scope, and business rules before accessing protected systems; the requirement instead turns on this behavior: explicit error signaling distinguishes tool failure from an empty successful result; trusting the caller's description is not a security control.

B:

The server supplies discoverable names, descriptions, and input schemas for the tool set; this cannot substitute for the behavior that explicit error signaling distinguishes tool failure from an empty successful result, because misleading descriptions can cause inappropriate model tool selection.

C:

A network transport supports remote requests under explicit authentication and deployment controls; in contrast, explicit error signaling distinguishes tool failure from an empty successful result, while opening an HTTP endpoint without access restrictions can expose powerful tools publicly.

Question 522

A key accidentally appears in a public repository commit and may have been copied. Which approach targets the deciding condition?

  1. Authenticate with a server-managed API credential
  2. Rotate exposed secrets immediately
  3. Separate environments and credentials
  4. Negotiate supported content types properly

Explanation

Correct Answer

B:

Credential revocation and replacement remove the leaked key's usefulness while investigators assess usage, subject to this requirement: deleting the visible commit does not guarantee copies of the secret are gone.

Incorrect Answers

A:

A trusted backend injects and safeguards the credential when making requests; in contrast, credential revocation and replacement remove the leaked key's usefulness while investigators assess usage, while embedding the key in front-end JavaScript makes it available to site visitors.

C:

Independent configuration and secrets reduce accidental cross-environment side effects; here, the controlling capability is that credential revocation and replacement remove the leaked key's usefulness while investigators assess usage, but sharing powerful production credentials with tests defeats environmental isolation.

D:

Content-type checks help the parser distinguish valid API data from an unexpected gateway body; the missing capability is more specific: credential revocation and replacement remove the leaked key's usefulness while investigators assess usage; blind JSON parsing obscures transport failures.

Question 523

Only one small service needs to check a narrow in-memory business rule. Which solution matches the actual constraint?

  1. Prefer a prompt template for pure language guidance
  2. Prefer API middleware for organization-wide policies
  3. Prefer one application-local function for a one-off feature
  4. Use an MCP server for a reusable integration

Explanation

Correct Answer

C:

A local function minimizes protocol, hosting, and dependency overhead, subject to this requirement: reusability infrastructure should be justified by real multiple-client requirements.

Incorrect Answers

A:

A maintained prompt template provides reusable text behavior without operating a tool; this cannot substitute for the behavior that a local function minimizes protocol, hosting, and dependency overhead, because creating an executable integration would add complexity to a text-only requirement.

B:

Shared middleware or gateway controls enforce cross-cutting service rules centrally; in contrast, a local function minimizes protocol, hosting, and dependency overhead, while an MCP server would not by itself govern unrelated direct inference requests.

D:

A protocol server exposes shared tools and resources with consistent discovery semantics; here, the controlling capability is that a local function minimizes protocol, hosting, and dependency overhead, but it adds infrastructure and identity concerns compared with a single local function.

Question 524

Several components disagree whether an order ID is numeric or a string. Which response is most defensible under these conditions?

  1. Extract the tool dispatcher into a tested module
  2. Introduce typed domain models
  3. Remove dead configuration flags after verification
  4. Use typed deserialization for a structured result

Explanation

Correct Answer

B:

Explicit types and validation prevent accidental incompatible assumptions at module boundaries; types cannot establish that an order actually belongs to the authorized user.

Incorrect Answers

A:

Separating dispatch logic creates a clear authorization and error-handling boundary; by comparison, the intended solution relies on this: explicit types and validation prevent accidental incompatible assumptions at module boundaries; rearranging code without tests can accidentally change tool semantics.

C:

Evidence of unused paths plus regression tests permit safe cleanup of obsolete logic, whereas the requirement calls for another behavior: explicit types and validation prevent accidental incompatible assumptions at module boundaries; deleting unknown switches without checking deployments can cause hidden failures.

D:

Typed parsing converts validated JSON into an explicit application data contract; however, the critical mechanism here is that explicit types and validation prevent accidental incompatible assumptions at module boundaries, and free-form text parsing with ad hoc regular expressions is brittle under wording changes.

Question 525

A logged-in user requests a refund through a tool but the order belongs to a different customer. Which response is most defensible under these conditions?

  1. Validate and execute the tool outside the model
  2. Validate caller authorization per invocation
  3. Require explicit user confirmation before expensive writes
  4. Allowlist tools per agent role

Explanation

Correct Answer

B:

The backend checks owner, resource, action, and policy before accepting the call; model-generated account identifiers are not an authorization proof.

Incorrect Answers

A:

Application code enforces permission checks and performs the actual side effect or read; by comparison, the intended solution relies on this: the backend checks owner, resource, action, and policy before accepting the call; the presence of tool_use content is not evidence the backend was queried.

C:

An enforced approval step separates informative recommendations from transactional execution, whereas the requirement calls for another behavior: the backend checks owner, resource, action, and policy before accepting the call; a prompt that says ask first cannot replace backend commit gating.

D:

Restricting registered capabilities limits possible operations at the dispatch boundary; however, the critical mechanism here is that the backend checks owner, resource, action, and policy before accepting the call, and a polite instruction cannot fully constrain a tool whose credentials are overprivileged.

Question 526

A chat product needs Claude to see relevant prior user and assistant turns for a follow-up. Which choice should be used here?

  1. Use the real-time Messages API for chat
  2. Use the top-level system instruction field
  3. Choose a supported model identifier
  4. Supply alternating conversation turns in messages

Explanation

Correct Answer

D:

Request messages carry explicit user and assistant roles because the API does not automatically retain prior chat state; the important constraint is that omitting history makes follow-up references ambiguous unless reintroduced elsewhere.

Incorrect Answers

A:

Direct Messages calls match interactive latency and incremental display requirements; by comparison, the intended solution relies on this: request messages carry explicit user and assistant roles because the API does not automatically retain prior chat state; batch jobs complete later and are inappropriate for a tightly coupled request-response interface.

B:

The Messages API represents governing instructions at request level rather than as a fabricated system role in messages; the needed capability is different—request messages carry explicit user and assistant roles because the API does not automatically retain prior chat state—and treating a system instruction as untrusted user content can blur instruction precedence.

C:

The model field identifies the deployed model; production should verify current IDs and compatibility, whereas the requirement calls for another behavior: request messages carry explicit user and assistant roles because the API does not automatically retain prior chat state; guessing an informal nickname can yield an invalid-model error or unexpected alias changes.

Question 527

A business user can modify a prompt template's tone but should not disable mandatory security checks. How should the implementation handle this need?

  1. Separate requirements from source data
  2. Separate policy from editable templates
  3. Validate prompt variables before interpolation
  4. Use a safe fallback for ambiguous instructions

Explanation

Correct Answer

B:

The application holds non-negotiable policies in protected configuration independent of user-edited prose; its operational boundary is that allowing end users to rewrite enforcement instructions weakens trust boundaries.

Incorrect Answers

A:

Structural boundaries distinguish data to analyze from commands the assistant should follow; this cannot substitute for the behavior that the application holds non-negotiable policies in protected configuration independent of user-edited prose, because external content should not gain authority merely because it appears near instructions.

C:

Input checks constrain length, types, and permitted resource scopes before prompt assembly; this addresses a separate concern, not the fact that the application holds non-negotiable policies in protected configuration independent of user-edited prose; merely surrounding an unsafe value with quotation marks does not establish authorization.

D:

The assistant asks for missing scope or routes the decision to an authorized operator; here, the controlling capability is that the application holds non-negotiable policies in protected configuration independent of user-edited prose, but guessing a target object could cause irreversible cross-customer damage.

Question 528

A network interruption occurs after a tool submitted a payment and the model may request it again. Which solution should the engineer adopt?

  1. Allowlist tools per agent role
  2. Use idempotency keys for retriable writes
  3. Include stable identifiers in tool results
  4. Require explicit user confirmation before expensive writes

Explanation

Correct Answer

B:

Stable transaction identifiers allow duplicate attempts to be detected before replaying the business action; a crucial qualification is that replaying without deduplication can create multiple charges.

Incorrect Answers

A:

Restricting registered capabilities limits possible operations at the dispatch boundary; here, the controlling capability is that stable transaction identifiers allow duplicate attempts to be detected before replaying the business action, but a polite instruction cannot fully constrain a tool whose credentials are overprivileged.

C:

IDs let downstream operations refer unambiguously to retrieved business objects; by comparison, the intended solution relies on this: stable transaction identifiers allow duplicate attempts to be detected before replaying the business action; human-readable names alone may collide across tenants or records.

D:

An enforced approval step separates informative recommendations from transactional execution; the needed capability is different—stable transaction identifiers allow duplicate attempts to be detected before replaying the business action—and a prompt that says ask first cannot replace backend commit gating.

Question 529

A diagnostic agent reads one tool result, adjusts its next investigation, and eventually returns a human-readable incident conclusion. Which TWO control-loop behaviors are needed at intermediate and final stages? Choose TWO.

  1. Finish with a user-facing answer after the last action
  2. Ask a clarifying question when critical input is missing
  3. Escalate after repeated inconclusive tool results
  4. Observe a tool result before choosing the next action
  5. Use LangGraph-style stateful orchestration

Explanation

Correct Answers

A:

The model synthesizes a final response once no additional client tool request is needed—with this limitation: a raw tool result is not automatically a contextualized customer answer.

D:

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; the observation determines whether another investigation step is warranted before the final answer.

Incorrect Answers

B:

Requesting an authorized scope prevents guesses from becoming unintended side effects; this cannot substitute for the behavior that the model synthesizes a final response once no additional client tool request is needed, because an internal plan cannot legitimately invent the missing tenancy information.

C:

Bounded fallback returns uncertainty and a trace for human follow-up rather than fabricating certainty; in contrast, the model synthesizes a final response once no additional client tool request is needed, while extra iterations are not justified once the cost and confidence guardrails trigger; an escalation is unnecessary once the agent can already produce the verified incident conclusion.

E:

A graph framework makes transitions and persisted workflow state first-class constructs; here, the controlling capability is that the model synthesizes a final response once no additional client tool request is needed, but the framework does not eliminate the need to secure the nodes' tools.

Question 530

An agent only needs to retrieve documents but its environment has broad administrative APIs. What is the correct engineering decision?

  1. Inspect outputs before taking business actions
  2. Apply input validation at ingress
  3. Use a policy-based refusal and safe redirect
  4. Use constrained tool schemas and allowlists

Explanation

Correct Answer

D:

Narrow tool capabilities reduce what even a manipulated model could attempt; its operational boundary is that a policy paragraph in system instructions is not a technical allowlist.

Incorrect Answers

A:

Output checks validate schema, scope, and safety before any backend mutation; however, the critical mechanism here is that narrow tool capabilities reduce what even a manipulated model could attempt, and post-generation checks cannot repair unauthorized access already performed upstream.

B:

Input controls reject invalid types, excessive sizes, and disallowed patterns before expensive processing; the requirement instead turns on this behavior: narrow tool capabilities reduce what even a manipulated model could attempt; content screening alone cannot establish the caller's resource permissions.

C:

A policy layer handles the request appropriately and communicates a safe boundary; this cannot substitute for the behavior that narrow tool capabilities reduce what even a manipulated model could attempt, because policy adherence should be evaluated; it is not guaranteed by a single warning prompt.

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