Anthropic CCDV-F (Claude Certified Developer - Foundations) Practice Test 15

Question 140

For a token-consumption optimization sprint, a design is acceptable only if it can serve high-volume straightforward classification at a tight latency and cost target while also being able to deliver moderately complex development assistance with balanced quality, cost, and latency. Which paired implementation is sufficient?

  1. Perform both steps: (1) Benchmark a suitable Haiku-family model for the simple task; (2) evaluate a sonnet-family model against a realistic task set.
  2. Perform both steps: (1) Select an available effort level after measuring task performance; (2) evaluate a sonnet-family model against a realistic task set.
  3. Perform both steps: (1) Benchmark a suitable Haiku-family model for the simple task; (2) track output tokens separately from input tokens.
  4. Perform both steps: (1) Select an available effort level after measuring task performance; (2) track output tokens separately from input tokens.

Explanation

Correct Answer

A:

The first change enables the team to serve high-volume straightforward classification at a tight latency and cost target because lightweight model tiers can favor speed and unit economics, but accuracy must still meet the specific evaluation threshold; the second addresses the separate need to deliver moderately complex development assistance with balanced quality, cost, and latency because a balanced model tier often offers useful capability without the latency or cost of the most expensive tier; evaluation should drive final selection.

Incorrect Answers

B:

The second step meets the requirement to deliver moderately complex development assistance with balanced quality, cost, and latency because a balanced model tier often offers useful capability without the latency or cost of the most expensive tier; evaluation should drive final selection; the first step would only help to trade answer latency and computation against solution quality on supported models because reasoning effort is an inference-time tradeoff; lowering it may reduce costs or latency while harming difficult-case accuracy, leaving the separate requirement to serve high-volume straightforward classification at a tight latency and cost target unmet.

C:

The first step is appropriate for the requirement to serve high-volume straightforward classification at a tight latency and cost target because lightweight model tiers can favor speed and unit economics, but accuracy must still meet the specific evaluation threshold; the other step instead aims to identify why verbose generated reports cost more despite unchanged inputs because generation tokens are metered independently and can have different pricing, so response verbosity affects unit economics, which does not meet the second requirement to deliver moderately complex development assistance with balanced quality, cost, and latency.

D:

Both steps target different results—trade answer latency and computation against solution quality on supported models and identify why verbose generated reports cost more despite unchanged inputs—because reasoning effort is an inference-time tradeoff; lowering it may reduce costs or latency while harming difficult-case accuracy and generation tokens are metered independently and can have different pricing, so response verbosity affects unit economics; the required capabilities are instead to serve high-volume straightforward classification at a tight latency and cost target and to deliver moderately complex development assistance with balanced quality, cost, and latency.

Question 141

For a production incident affecting a Claude endpoint, a design is acceptable only if it can explain why an agent loop is stuck after a model requested an external function while also being able to locate which component of a multi-tool answer introduced the wrong value. Which paired implementation is sufficient?

  1. Perform both steps: (1) Check that each tool_use identifier receives one matching tool_result; (2) read the 400 error details and validate the request payload.
  2. Perform both steps: (1) Inspect the HTTP 429 status and rate-limit metadata before retrying; (2) read the 400 error details and validate the request payload.
  3. Perform both steps: (1) Check that each tool_use identifier receives one matching tool_result; (2) examine correlated request, tool, and transformation traces stage by stage.
  4. Perform both steps: (1) Inspect the HTTP 429 status and rate-limit metadata before retrying; (2) examine correlated request, tool, and transformation traces stage by stage.

Explanation

Correct Answer

C:

The first change enables the team to explain why an agent loop is stuck after a model requested an external function because an unmatched tool call leaves a protocol-level continuation gap, even if the handler actually ran successfully; the second addresses the separate need to locate which component of a multi-tool answer introduced the wrong value because structured traces reveal whether retrieval, tool execution, postprocessing, or generation caused the defect; reading only the final text cannot isolate it.

Incorrect Answers

A:

The first step is appropriate for the requirement to explain why an agent loop is stuck after a model requested an external function because an unmatched tool call leaves a protocol-level continuation gap, even if the handler actually ran successfully; the other step instead aims to find a request schema error instead of repeatedly changing the system prompt because an invalid API field or incompatible parameter cannot be fixed by natural-language instructions within the malformed request, which does not meet the second requirement to locate which component of a multi-tool answer introduced the wrong value.

B:

Both steps target different results—distinguish an upstream throttling response from a malformed model answer and find a request schema error instead of repeatedly changing the system prompt—because a rate limit is a request-layer capacity problem; the response needs controlled pacing rather than a prompt rewrite and an invalid API field or incompatible parameter cannot be fixed by natural-language instructions within the malformed request; the required capabilities are instead to explain why an agent loop is stuck after a model requested an external function and to locate which component of a multi-tool answer introduced the wrong value.

D:

The second step meets the requirement to locate which component of a multi-tool answer introduced the wrong value because structured traces reveal whether retrieval, tool execution, postprocessing, or generation caused the defect; reading only the final text cannot isolate it; the first step would only help to distinguish an upstream throttling response from a malformed model answer because a rate limit is a request-layer capacity problem; the response needs controlled pacing rather than a prompt rewrite, leaving the separate requirement to explain why an agent loop is stuck after a model requested an external function unmet.

Question 142

An operational review of a secure tool-execution boundary reports distinct failures: it cannot avoid accidentally allowing critical writes when a safety check crashes, and it cannot keep Claude API credentials out of repository source and deployment images. Which two repairs are required? Choose TWO.

  1. Combine prompt boundaries, scoped credentials, validation, and approval gates.
  2. Use layered policy enforcement and restrict powerful tools independently of chat text.
  3. Define fail-closed behavior for an enforcement hook and test failure paths.
  4. Validate or moderate final content before release.
  5. Retrieve keys through a managed secrets facility at runtime.

Explanation

Correct Answers

C:

To avoid accidentally allowing critical writes when a safety check crashes as part of a design that must also keep Claude API credentials out of repository source and deployment images, a guardrail that silently skips checks on errors is not an effective boundary; failures must prevent the protected action when policy requires it; this is why the action is one of the two required selections.

E:

To keep Claude API credentials out of repository source and deployment images as part of a design that must also avoid accidentally allowing critical writes when a safety check crashes, central secret handling improves rotation and auditing and avoids accidental credential disclosure through source control; this is why the action is one of the two required selections.

Incorrect Answers

A:

This alternative serves the separate goal to secure an agent against both malicious retrieved text and unsafe tool execution because defense in depth assumes individual filters can fail; independent enforcement boundaries reduce the chance of harmful side effects; the required outcomes are instead to avoid accidentally allowing critical writes when a safety check crashes and to keep Claude API credentials out of repository source and deployment images.

B:

This alternative serves the separate goal to prevent an adversarial user from bypassing safety policy by roleplay instructions because adversarial prompting can target instruction hierarchy, so product controls and tool authorization must remain effective when wording is manipulated; the required outcomes are instead to avoid accidentally allowing critical writes when a safety check crashes and to keep Claude API credentials out of repository source and deployment images.

D:

This alternative serves the separate goal to prevent generated advice that violates the product's publication rules from being shown because output checks catch issues emerging during generation; they complement rather than replace input controls and tool restrictions; the required outcomes are instead to avoid accidentally allowing critical writes when a safety check crashes and to keep Claude API credentials out of repository source and deployment images.

Question 143

While planning a function-calling service, the product owner specifies that it must process several requested independent tools without mixing their outputs. Which solution fulfills the requirement?

  1. Configure the built-in capability instead of recreating it as a local handler.
  2. Correlate each tool result with its matching tool_use_id.
  3. Package the connector behind an MCP server.
  4. Put an approval gate in the host-side write handler.

Explanation

Correct Answer

B:

To process several requested independent tools without mixing their outputs in this function-calling service, multi-tool responses can interleave operations, so identifier-based correlation is essential for safe continuation; the proposed implementation directly supplies that capability.

Incorrect Answers

A:

The proposed alternative helps to obtain a capability already provided as a supported Anthropic server tool because provider tools simplify execution ownership, though availability, cost, and data handling still require assessment; that mechanism does not meet the different requirement to process several requested independent tools without mixing their outputs.

C:

The proposed alternative helps to share one integration with several compatible AI clients or agent runtimes because MCP standardizes capability exposure across clients, reducing duplicate integration logic compared with proprietary tool schemas everywhere; that mechanism does not meet the different requirement to process several requested independent tools without mixing their outputs.

D:

The proposed alternative helps to ensure an agent obtains explicit authorization before updating a customer record because an irreversible or sensitive write must be permitted by trusted application logic, not merely because the model selected a tool; that mechanism does not meet the different requirement to process several requested independent tools without mixing their outputs.

Question 144

The owner of a production inference application needs to reuse a long stable instruction or document prefix over repeated requests as well as explain why a previously cached prompt is now charged as new input. Which two changes belong in the minimum remediation plan? Choose TWO.

  1. Use stop_sequences and inspect the returned stop_sequence field.
  2. Parse response content by block type and follow current thinking-field rules.
  3. Cache an eligible unchanged prefix with cache_control.
  4. Inspect changes to content and ordering before the cache breakpoint.
  5. Assemble messages with correct user and assistant roles.

Explanation

Correct Answers

C:

To reuse a long stable instruction or document prefix over repeated requests as part of a design that must also explain why a previously cached prompt is now charged as new input, prompt caching reduces repeated input processing when the prefix matches; changes before the cache boundary can prevent a hit; this is why the action is one of the two required selections.

D:

To explain why a previously cached prompt is now charged as new input as part of a design that must also reuse a long stable instruction or document prefix over repeated requests, cache reuse depends on matching prefix material; adding text to the front changes what is cached even if later text is unchanged; this is why the action is one of the two required selections.

Incorrect Answers

A:

This alternative serves the separate goal to enforce a caller-defined textual boundary on generation because a stop sequence halts generation at a specified marker, distinct from natural completion and separate from a JSON schema validator; the required outcomes are instead to reuse a long stable instruction or document prefix over repeated requests and to explain why a previously cached prompt is now charged as new input.

B:

This alternative serves the separate goal to handle a response containing reasoning-related blocks without confusing them with ordinary final text because structured content may contain distinct block types and signatures; a parser that assumes only plain text can discard important protocol data; the required outcomes are instead to reuse a long stable instruction or document prefix over repeated requests and to explain why a previously cached prompt is now charged as new input.

E:

This alternative serves the separate goal to distinguish an end user's request from previous assistant turns in a dialogue because conversation structure influences how the model interprets dialogue and tool history; flattening roles into one text string loses that structure; the required outcomes are instead to reuse a long stable instruction or document prefix over repeated requests and to explain why a previously cached prompt is now charged as new input.

Question 145

The owner of a document investigation agent needs to speed up independent data lookups without causing write-order races as well as pass reliable machine-readable outputs between independently developed agent stages. Which two changes belong in the minimum remediation plan? Choose TWO.

  1. Prefer a simple API call over a general-purpose multi-agent framework.
  2. Run side-effect-free tool calls concurrently when their inputs are independent.
  3. Define and validate a typed schema at the handoff.
  4. Attach a post-execution observation hook to the tool lifecycle.
  5. Choose a graph-based orchestration framework when state transitions justify it.

Explanation

Correct Answers

B:

To speed up independent data lookups without causing write-order races as part of a design that must also pass reliable machine-readable outputs between independently developed agent stages, read-only independent calls can safely overlap; mutating operations need explicit ordering or concurrency control; this is why the action is one of the two required selections.

C:

To pass reliable machine-readable outputs between independently developed agent stages as part of a design that must also speed up independent data lookups without causing write-order races, typed outputs make downstream contracts inspectable and reduce ambiguity caused by free-form prose passed as control data; this is why the action is one of the two required selections.

Incorrect Answers

A:

This alternative serves the separate goal to avoid adding orchestration infrastructure to a single request-and-response feature because frameworks are useful for genuinely branching, long-lived execution, but increase complexity when no orchestration problem exists; the required outcomes are instead to speed up independent data lookups without causing write-order races and to pass reliable machine-readable outputs between independently developed agent stages.

D:

This alternative serves the separate goal to inspect the result of a completed tool call before later model reasoning consumes it because post-tool inspection can record or validate outcomes after execution, but it cannot retroactively prevent side effects from the completed call; the required outcomes are instead to speed up independent data lookups without causing write-order races and to pass reliable machine-readable outputs between independently developed agent stages.

E:

This alternative serves the separate goal to coordinate explicit graph transitions among multiple model and tool nodes because graph orchestration clarifies conditional edges and resumability; it adds complexity when a short linear function would suffice; the required outcomes are instead to speed up independent data lookups without causing write-order races and to pass reliable machine-readable outputs between independently developed agent stages.

Question 146

Two acceptance tests block an MCP integration server: one requires the ability to enforce a mandatory fixed approval step regardless of what instructions the model follows; the other requires the ability to make a repeatable multi-step code-review methodology available to an agent. Which plan clears both tests?

  1. Perform both steps: (1) Implement the approval in application workflow control rather than only in a Skill; (2) create a skill containing the procedure and supporting instructions.
  2. Perform both steps: (1) Implement the appropriate MCP capability listing and metadata endpoints; (2) return a concise task-relevant projection from the tool handler.
  3. Perform both steps: (1) Implement the approval in application workflow control rather than only in a Skill; (2) return a concise task-relevant projection from the tool handler.
  4. Perform both steps: (1) Implement the appropriate MCP capability listing and metadata endpoints; (2) create a skill containing the procedure and supporting instructions.

Explanation

Correct Answer

A:

The first change enables the team to enforce a mandatory fixed approval step regardless of what instructions the model follows because skills can guide agent behavior, but deterministic policy gates must live in trusted code so they cannot be skipped by model interpretation; the second addresses the separate need to make a repeatable multi-step code-review methodology available to an agent because a Skill teaches how to perform a task, whereas a tool is the executable capability that performs an external operation.

Incorrect Answers

B:

Both steps target different results—help a client learn which callable and informational capabilities a server offers and reduce context bloat caused by a tool returning a thousand irrelevant fields—because discovery advertises tools, resources, or prompts so clients can use supported contracts rather than guessing names or arguments and tool output is subsequently consumed as model context; minimizing it reduces both token cost and distractions; the required capabilities are instead to enforce a mandatory fixed approval step regardless of what instructions the model follows and to make a repeatable multi-step code-review methodology available to an agent.

C:

The first step is appropriate for the requirement to enforce a mandatory fixed approval step regardless of what instructions the model follows because skills can guide agent behavior, but deterministic policy gates must live in trusted code so they cannot be skipped by model interpretation; the other step instead aims to reduce context bloat caused by a tool returning a thousand irrelevant fields because tool output is subsequently consumed as model context; minimizing it reduces both token cost and distractions, which does not meet the second requirement to make a repeatable multi-step code-review methodology available to an agent.

D:

The second step meets the requirement to make a repeatable multi-step code-review methodology available to an agent because a Skill teaches how to perform a task, whereas a tool is the executable capability that performs an external operation; the first step would only help to help a client learn which callable and informational capabilities a server offers because discovery advertises tools, resources, or prompts so clients can use supported contracts rather than guessing names or arguments, leaving the separate requirement to enforce a mandatory fixed approval step regardless of what instructions the model follows unmet.

Question 147

A team maintaining a Claude API workload benchmark has two different engineering requirements: (1) detect a looping agent that suddenly multiplies token consumption, and (2) forecast spend for a batch of documents with different input sizes. Which TWO independent changes directly meet them? Choose TWO.

  1. Compare quality-adjusted cost per accepted result across candidate models.
  2. Report p50, p95, and p99 latency by model and operation.
  3. Alert on unusual per-job token and tool-call growth.
  4. Prefer a maintained SDK when its typed methods and retry behavior fit the integration.
  5. Estimate cost from expected input and output tokens and current model rates.

Explanation

Correct Answers

C:

To detect a looping agent that suddenly multiplies token consumption as part of a design that must also forecast spend for a batch of documents with different input sizes, monitoring consumption across an entire agent run catches runaway workflows that per-request logs may not reveal; this is why the action is one of the two required selections.

E:

To forecast spend for a batch of documents with different input sizes as part of a design that must also detect a looping agent that suddenly multiplies token consumption, workload cost depends on token mix, model choice, and any applicable discounts or caching rather than request count alone; this is why the action is one of the two required selections.

Incorrect Answers

A:

This alternative serves the separate goal to choose a cheaper model only when the workload-level economics improve because a lower token price can be offset by retries, higher failure rates, or excessive output, so accepted-result cost is the useful measure; the required outcomes are instead to detect a looping agent that suddenly multiplies token consumption and to forecast spend for a batch of documents with different input sizes.

B:

This alternative serves the separate goal to avoid overlooking slow tail requests hidden by a good average because percentiles expose user experience at the tail and can reveal bottlenecks that a mean response time conceals; the required outcomes are instead to detect a looping agent that suddenly multiplies token consumption and to forecast spend for a batch of documents with different input sizes.

D:

This alternative serves the separate goal to choose between writing raw REST requests and using an official client library because sDKs wrap the same underlying service API and can simplify serialization, error types, and streaming without removing protocol constraints; the required outcomes are instead to detect a looping agent that suddenly multiplies token consumption and to forecast spend for a batch of documents with different input sizes.

Question 148

A production readiness review of a specialist-agent orchestration service finds that the team cannot yet deliver intermediate agent events to the client as they occur or restrict a development agent to a defined set of file and command capabilities. Which paired changes directly remedy the two gaps?

  1. Perform both steps: (1) Define and validate a typed schema at the handoff; (2) configure the appropriate anthropic-executed server tool.
  2. Perform both steps: (1) Consume the agent runtime's streamed event sequence; (2) configure tool permissions rather than depending only on a prompt.
  3. Perform both steps: (1) Define and validate a typed schema at the handoff; (2) configure tool permissions rather than depending only on a prompt.
  4. Perform both steps: (1) Consume the agent runtime's streamed event sequence; (2) configure the appropriate anthropic-executed server tool.

Explanation

Correct Answer

B:

The first change enables the team to deliver intermediate agent events to the client as they occur because event-driven consumption lets the application display progress and inspect tool actions without waiting for the entire multi-step run to finish; the second addresses the separate need to restrict a development agent to a defined set of file and command capabilities because runtime permissions constrain what actions can run even if the model requests more access; prompt instructions alone are not a security boundary.

Incorrect Answers

A:

Both steps target different results—pass reliable machine-readable outputs between independently developed agent stages and use a Claude platform-provided capability without operating a local handler—because typed outputs make downstream contracts inspectable and reduce ambiguity caused by free-form prose passed as control data and server tools execute on the provider side and have distinct authorization and runtime semantics from client tools the application dispatches; the required capabilities are instead to deliver intermediate agent events to the client as they occur and to restrict a development agent to a defined set of file and command capabilities.

C:

The second step meets the requirement to restrict a development agent to a defined set of file and command capabilities because runtime permissions constrain what actions can run even if the model requests more access; prompt instructions alone are not a security boundary; the first step would only help to pass reliable machine-readable outputs between independently developed agent stages because typed outputs make downstream contracts inspectable and reduce ambiguity caused by free-form prose passed as control data, leaving the separate requirement to deliver intermediate agent events to the client as they occur unmet.

D:

The first step is appropriate for the requirement to deliver intermediate agent events to the client as they occur because event-driven consumption lets the application display progress and inspect tool actions without waiting for the entire multi-step run to finish; the other step instead aims to use a Claude platform-provided capability without operating a local handler because server tools execute on the provider side and have distinct authorization and runtime semantics from client tools the application dispatches, which does not meet the second requirement to restrict a development agent to a defined set of file and command capabilities.

Question 149

An end-to-end test for a multi-tenant support application checks whether it can keep sensitive case material within approved processing and retention boundaries and identify damage that could result from adversarial documents or tool access. Which plan should be approved to satisfy both checks?

  1. Perform both steps: (1) Store team settings in the project's shared configuration file; (2) create a threat model covering trust boundaries, permissions, and misuse paths.
  2. Perform both steps: (1) Document permitted processing regions and data-handling obligations before selecting a deployment; (2) authenticate incoming callbacks and validate their signatures where supported.
  3. Perform both steps: (1) Document permitted processing regions and data-handling obligations before selecting a deployment; (2) create a threat model covering trust boundaries, permissions, and misuse paths.
  4. Perform both steps: (1) Store team settings in the project's shared configuration file; (2) authenticate incoming callbacks and validate their signatures where supported.

Explanation

Correct Answer

C:

The first change enables the team to keep sensitive case material within approved processing and retention boundaries because legal or organizational constraints must inform provider routing, logging, and retention rather than being added after data already flows; the second addresses the separate need to identify damage that could result from adversarial documents or tool access because aI application risk includes injected content and powerful downstream functions; those flows need explicit mitigations before implementation.

Incorrect Answers

A:

The second step meets the requirement to identify damage that could result from adversarial documents or tool access because aI application risk includes injected content and powerful downstream functions; those flows need explicit mitigations before implementation; the first step would only help to apply consistent reviewed tool permissions to an entire Claude Code repository team because repository-level settings can be versioned and reviewed, unlike private preferences that vary across each developer's machine, leaving the separate requirement to keep sensitive case material within approved processing and retention boundaries unmet.

B:

The first step is appropriate for the requirement to keep sensitive case material within approved processing and retention boundaries because legal or organizational constraints must inform provider routing, logging, and retention rather than being added after data already flows; the other step instead aims to trust a callback only after verifying who actually sent it because network arrival alone does not prove a webhook's origin; authenticity checks prevent forged events from changing agent state, which does not meet the second requirement to identify damage that could result from adversarial documents or tool access.

D:

Both steps target different results—apply consistent reviewed tool permissions to an entire Claude Code repository team and trust a callback only after verifying who actually sent it—because repository-level settings can be versioned and reviewed, unlike private preferences that vary across each developer's machine and network arrival alone does not prove a webhook's origin; authenticity checks prevent forged events from changing agent state; the required capabilities are instead to keep sensitive case material within approved processing and retention boundaries and to identify damage that could result from adversarial documents or tool access.

Question 150

To ship an internal agent-extension rollout, the application must stop a slow external system from trapping an agent loop indefinitely and connect a Claude tool request to an internal service whose logic the application owns. Which pair of listed actions is necessary for those stated capabilities? Choose TWO.

  1. Return a concise task-relevant projection from the tool handler.
  2. Implement a client-side handler and return its result with the tool-use identifier.
  3. Return an explicit tool error indication and an informative safe message.
  4. Give each tool a timeout and controlled retry or failure result.
  5. Define a descriptive tool name and validated input schema.

Explanation

Correct Answers

B:

To connect a Claude tool request to an internal service whose logic the application owns as part of a design that must also stop a slow external system from trapping an agent loop indefinitely, a declared client tool does not execute itself; application code runs it and feeds results back into the conversation; this is why the action is one of the two required selections.

D:

To stop a slow external system from trapping an agent loop indefinitely as part of a design that must also connect a Claude tool request to an internal service whose logic the application owns, tool execution is a separate latency source; bounded calls prevent one dependency from exhausting the entire agent budget; this is why the action is one of the two required selections.

Incorrect Answers

A:

This alternative serves the separate goal to reduce context bloat caused by a tool returning a thousand irrelevant fields because tool output is subsequently consumed as model context; minimizing it reduces both token cost and distractions; the required outcomes are instead to stop a slow external system from trapping an agent loop indefinitely and to connect a Claude tool request to an internal service whose logic the application owns.

C:

This alternative serves the separate goal to let Claude distinguish a failed database lookup from an empty successful result because an error result preserves failure semantics, preventing the model from inventing a no-record finding for a call that never succeeded; the required outcomes are instead to stop a slow external system from trapping an agent loop indefinitely and to connect a Claude tool request to an internal service whose logic the application owns.

E:

This alternative serves the separate goal to expose a business lookup with predictable typed arguments because well-described schemas help Claude select a capability and emit usable arguments; the backend still validates authorization and meaning; the required outcomes are instead to stop a slow external system from trapping an agent loop indefinitely and to connect a Claude tool request to an internal service whose logic the application owns.

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