Anthropic CCDV-F (Claude Certified Developer - Foundations) Practice Test 23

Question 220

A language-model integration review asks the team to close two unrelated gaps without merging their execution paths. Service A must distinguish borderline policy labels with examples of both allowed and disallowed cases; service B must avoid assuming retrieved material, tool outputs, history, and instructions each have independent infinite capacity. Which paired remediation works?

  1. First component: Use contrasting few-shot examples. Second component: Understand context-window competition.
  2. First component: Reduce reasoning effort on trivial routing. Second component: Understand context-window competition.
  3. First component: Use one representative example. Second component: Reduce reasoning effort on trivial routing.
  4. First component: Use contrasting few-shot examples. Second component: Use one representative example.

Explanation

Correct Answer

A:

The first proposal fits its assigned role because multiple representative cases help define a decision boundary and reduce ambiguity, while the second proposal fits its assigned role because all supplied content competes for a finite effective context budget and may affect attention; their respective acceptance criteria are to distinguish borderline policy labels with examples of both allowed and disallowed cases and to avoid assuming retrieved material, tool outputs, history, and instructions each have independent infinite capacity.

Incorrect Answers

B:

The first proposal serves the different purpose to shorten latency and spend on tasks that need only basic categorization instead of the assigned first role because lower effort can suit predictable work where deeper reasoning adds little measurable quality, while the second proposal fits its assigned role because all supplied content competes for a finite effective context budget and may affect attention; their respective acceptance criteria are to distinguish borderline policy labels with examples of both allowed and disallowed cases and to avoid assuming retrieved material, tool outputs, history, and instructions each have independent infinite capacity.

C:

The first proposal serves the different purpose to clarify a specific formatting convention that a plain instruction was not conveying instead of the assigned first role because a single demonstration can anchor the expected shape and tone of a generated response, while the second proposal serves the different purpose to shorten latency and spend on tasks that need only basic categorization instead of the assigned second role because lower effort can suit predictable work where deeper reasoning adds little measurable quality; their respective acceptance criteria are to distinguish borderline policy labels with examples of both allowed and disallowed cases and to avoid assuming retrieved material, tool outputs, history, and instructions each have independent infinite capacity.

D:

The first proposal fits its assigned role because multiple representative cases help define a decision boundary and reduce ambiguity, while the second proposal serves the different purpose to clarify a specific formatting convention that a plain instruction was not conveying instead of the assigned second role because a single demonstration can anchor the expected shape and tone of a generated response; their respective acceptance criteria are to distinguish borderline policy labels with examples of both allowed and disallowed cases and to avoid assuming retrieved material, tool outputs, history, and instructions each have independent infinite capacity.

Question 221

A release plan for a model platform platform has two acceptance checks: component one must serve a mixed coding and support assistant at moderate latency and cost; component two must choose between releases using an actual organization's error and latency measurements. Which proposal passes both?

  1. First component: Prefer a Sonnet-class model for balanced workloads. Second component: Prefer a Haiku-class model for lightweight routing.
  2. First component: Prefer a Haiku-class model for lightweight routing. Second component: Prefer a high-capability Opus-class model.
  3. First component: Prefer a Sonnet-class model for balanced workloads. Second component: Run task-specific model evaluations.
  4. First component: Prefer a high-capability Opus-class model. Second component: Run task-specific model evaluations.

Explanation

Correct Answer

C:

The first proposal fits its assigned role because the balanced tier often offers a practical quality-throughput compromise, subject to task-level testing, while the second proposal fits its assigned role because representative tests expose differences hidden by generic capability labels; their respective acceptance criteria are to serve a mixed coding and support assistant at moderate latency and cost and to choose between releases using an actual organization's error and latency measurements.

Incorrect Answers

A:

The first proposal fits its assigned role because the balanced tier often offers a practical quality-throughput compromise, subject to task-level testing, while the second proposal serves the different purpose to classify many short messages where low latency and unit cost dominate instead of the assigned second role because the lighter tier suits inexpensive, high-throughput tasks with modest reasoning demand; their respective acceptance criteria are to serve a mixed coding and support assistant at moderate latency and cost and to choose between releases using an actual organization's error and latency measurements.

B:

The first proposal serves the different purpose to classify many short messages where low latency and unit cost dominate instead of the assigned first role because the lighter tier suits inexpensive, high-throughput tasks with modest reasoning demand, while the second proposal serves the different purpose to handle a rare complex architecture review where reasoning accuracy outweighs cost and latency instead of the assigned second role because the capability-oriented tier is suited to the hardest tasks after evaluation confirms an advantage; their respective acceptance criteria are to serve a mixed coding and support assistant at moderate latency and cost and to choose between releases using an actual organization's error and latency measurements.

D:

The first proposal serves the different purpose to handle a rare complex architecture review where reasoning accuracy outweighs cost and latency instead of the assigned first role because the capability-oriented tier is suited to the hardest tasks after evaluation confirms an advantage, while the second proposal fits its assigned role because representative tests expose differences hidden by generic capability labels; their respective acceptance criteria are to serve a mixed coding and support assistant at moderate latency and cost and to choose between releases using an actual organization's error and latency measurements.

Question 222

An output validation application is approaching release, but the team still needs to keep an embedded webpage instruction from becoming higher-priority application policy. What is the most appropriate implementation?

  1. Anchor summaries to source identifiers
  2. Prune obsolete tool outputs
  3. Reserve space for the model's output
  4. Preserve instruction and evidence boundaries

Explanation

Correct Answer

D:

Context assembly must label external text as data and avoid granting it authority over tool use; this behavior directly supports the need to keep an embedded webpage instruction from becoming higher-priority application policy. Quotation formatting alone does not enforce sandboxing or authorization.

Incorrect Answers

A:

Stable document IDs and passage references preserve provenance beyond the shortened working memory; this applies to the different goal to enable a reviewer to reconstruct evidence after context has been compacted, not the requirement to keep an embedded webpage instruction from becoming higher-priority application policy; a bare compressed claim cannot show which source established it.

B:

When an application needs to stop stale log dumps from crowding out new evidence during an agent session, tool-result pruning removes redundant bulky observations after their findings are retained; the current capability must instead keep an embedded webpage instruction from becoming higher-priority application policy, while keeping all raw logs forever does not guarantee better reasoning.

C:

Input assembly must leave room for the requested completion and any needed reasoning; the supported outcome is to avoid using the entire context budget on retrieved input before generating a response, whereas this scenario requires the ability to keep an embedded webpage instruction from becoming higher-priority application policy; a technically accepted long prompt can still prevent a useful-length answer.

Question 223

After pilot feedback, a conversation design group needs two independent improvements. The first is to improve extraction quality after discovering consistent edge-case failures; the second is to distinguish near-duplicate support intents that a broad rule routinely confuses. Which engineering proposal handles both?

  1. First component: Iterate prompts against an evaluation set. Second component: Use explicit refusal and escalation instructions.
  2. First component: Specify output format in the prompt. Second component: Use few-shot boundary cases.
  3. First component: Iterate prompts against an evaluation set. Second component: Use few-shot boundary cases.
  4. First component: Use explicit refusal and escalation instructions. Second component: Specify output format in the prompt.

Explanation

Correct Answer

C:

The first proposal fits its assigned role because controlled revisions measured on stable examples show whether a change actually helps, while the second proposal fits its assigned role because examples illustrating each side of a decision boundary communicate nuanced classification criteria; their respective acceptance criteria are to improve extraction quality after discovering consistent edge-case failures and to distinguish near-duplicate support intents that a broad rule routinely confuses.

Incorrect Answers

A:

The first proposal fits its assigned role because controlled revisions measured on stable examples show whether a change actually helps, while the second proposal serves the different purpose to prevent a customer-facing assistant from making final legal determinations without review instead of the assigned second role because clear boundaries describe when to abstain or route to a qualified person; their respective acceptance criteria are to improve extraction quality after discovering consistent edge-case failures and to distinguish near-duplicate support intents that a broad rule routinely confuses.

B:

The first proposal serves the different purpose to obtain named fields in a consistent response layout for subsequent processing instead of the assigned first role because explicit field descriptions guide generation toward the consumer's expected structure, while the second proposal fits its assigned role because examples illustrating each side of a decision boundary communicate nuanced classification criteria; their respective acceptance criteria are to improve extraction quality after discovering consistent edge-case failures and to distinguish near-duplicate support intents that a broad rule routinely confuses.

D:

The first proposal serves the different purpose to prevent a customer-facing assistant from making final legal determinations without review instead of the assigned first role because clear boundaries describe when to abstain or route to a qualified person, while the second proposal serves the different purpose to obtain named fields in a consistent response layout for subsequent processing instead of the assigned second role because explicit field descriptions guide generation toward the consumer's expected structure; their respective acceptance criteria are to improve extraction quality after discovering consistent edge-case failures and to distinguish near-duplicate support intents that a broad rule routinely confuses.

Question 224

A developer is reviewing alternatives for an inference routing feature that must reduce token spend without losing facts needed for the current request. Which option is a defensible fit?

  1. Keep volatile fields after reusable context
  2. Shorten irrelevant conversation history
  3. Use model routing for simple versus hard cases
  4. Estimate output length before scaling
  5. Use batch pricing where deadline permits

Explanation

Correct Answer

B:

Selective context assembly preserves active constraints while removing obsolete turns; this behavior directly supports the need to reduce token spend without losing facts needed for the current request. Arbitrary truncation can erase information needed for correct action.

Incorrect Answers

A:

Stable instruction and reference sections should precede changing per-request details for effective prefix reuse; despite being useful when a service must prevent request timestamps from invalidating a costly static prefix, it is not a solution to reduce token spend without losing facts needed for the current request; placing a changing value before the cache boundary undermines matching.

C:

Conditional routing improves blended cost without discarding an escalation path; this applies to the different goal to spend high-capability tokens only when validation indicates complexity, not the requirement to reduce token spend without losing facts needed for the current request; blindly downgrading every call can increase retries and review expenses.

D:

When an application needs to avoid underbudgeting a report-writing feature whose completions are much longer than inputs, cost models should include the generated token volume and task-specific variance; the current capability must instead reduce token spend without losing facts needed for the current request, while input-only estimates miss potentially dominant completion expense.

E:

Asynchronous batch processing can offer a cost-effective throughput path under supported terms; the supported outcome is to lower spend for large scheduled classification jobs that do not need immediate responses, whereas this scenario requires the ability to reduce token spend without losing facts needed for the current request; a night job should not be optimized with live streaming unless its users need it.

Question 225

A token economics review requires two independent, explicitly defined outcomes: the first workstream must control average spend while retaining quality on exceptional requests; a second must serve a mixed coding and support assistant at moderate latency and cost. Which TWO actions directly provide these capabilities? Choose TWO.

  1. Route hard cases to a stronger fallback
  2. Check modality and feature support
  3. Measure end-to-end latency, not just token speed
  4. Prefer a Sonnet-class model for balanced workloads
  5. Prefer a high-capability Opus-class model

Explanation

Correct Answers

A:

A tiered policy starts with a cheaper path and escalates on measurable uncertainty or validation failure; this directly supports task 1 to control average spend while retaining quality on exceptional requests, not the separate task or tasks to serve a mixed coding and support assistant at moderate latency and cost; fixed use of the most costly model discards easy-case savings.

D:

The balanced tier often offers a practical quality-throughput compromise, subject to task-level testing; this directly supports task 2 to serve a mixed coding and support assistant at moderate latency and cost, not the separate task or tasks to control average spend while retaining quality on exceptional requests; it should not be assumed to dominate every model on every benchmark.

Incorrect Answers

B:

Availability of context length, structured output, and modalities may depend on the selected model and version; this implements the different outcome to avoid selecting a model that lacks a required document or tool feature rather than the required tasks to serve a mixed coding and support assistant at moderate latency and cost and to control average spend while retaining quality on exceptional requests, and family-level generalizations do not override specific compatibility documentation, so it is not a direct solution to any of the assigned workstreams.

C:

Total response time includes queuing, input processing, generation, and downstream integration; this implements the different outcome to select the right model for an application with tool and network overhead rather than the required tasks to serve a mixed coding and support assistant at moderate latency and cost and to control average spend while retaining quality on exceptional requests, and a high tokens-per-second figure alone can conceal slow first-token delivery, so it is not a direct solution to any of the assigned workstreams.

E:

The capability-oriented tier is suited to the hardest tasks after evaluation confirms an advantage; this implements the different outcome to handle a rare complex architecture review where reasoning accuracy outweighs cost and latency rather than the required tasks to serve a mixed coding and support assistant at moderate latency and cost and to control average spend while retaining quality on exceptional requests, and it is an inefficient default for every basic classification request, so it is not a direct solution to any of the assigned workstreams.

Question 226

Engineers are splitting a prompt engineering service into two independently tested functions. Function A must enable a reviewer to reconstruct evidence after context has been compacted, while function B must stop stale log dumps from crowding out new evidence during an agent session. Which plan is sound?

  1. First component: Anchor summaries to source identifiers. Second component: Prune obsolete tool outputs.
  2. First component: Use scoped subagent contexts. Second component: Reserve space for the model's output.
  3. First component: Reserve space for the model's output. Second component: Prune obsolete tool outputs.
  4. First component: Anchor summaries to source identifiers. Second component: Use scoped subagent contexts.

Explanation

Correct Answer

A:

The first proposal fits its assigned role because stable document IDs and passage references preserve provenance beyond the shortened working memory, while the second proposal fits its assigned role because tool-result pruning removes redundant bulky observations after their findings are retained; their respective acceptance criteria are to enable a reviewer to reconstruct evidence after context has been compacted and to stop stale log dumps from crowding out new evidence during an agent session.

Incorrect Answers

B:

The first proposal serves the different purpose to prevent a specialist's exhaustive document scan from contaminating the coordinator's working context instead of the assigned first role because isolated workers return concise findings while retaining intermediate detail within their own window, while the second proposal serves the different purpose to avoid using the entire context budget on retrieved input before generating a response instead of the assigned second role because input assembly must leave room for the requested completion and any needed reasoning; their respective acceptance criteria are to enable a reviewer to reconstruct evidence after context has been compacted and to stop stale log dumps from crowding out new evidence during an agent session.

C:

The first proposal serves the different purpose to avoid using the entire context budget on retrieved input before generating a response instead of the assigned first role because input assembly must leave room for the requested completion and any needed reasoning, while the second proposal fits its assigned role because tool-result pruning removes redundant bulky observations after their findings are retained; their respective acceptance criteria are to enable a reviewer to reconstruct evidence after context has been compacted and to stop stale log dumps from crowding out new evidence during an agent session.

D:

The first proposal fits its assigned role because stable document IDs and passage references preserve provenance beyond the shortened working memory, while the second proposal serves the different purpose to prevent a specialist's exhaustive document scan from contaminating the coordinator's working context instead of the assigned second role because isolated workers return concise findings while retaining intermediate detail within their own window; their respective acceptance criteria are to enable a reviewer to reconstruct evidence after context has been compacted and to stop stale log dumps from crowding out new evidence during an agent session.

Question 227

A production-readiness check for a customer platform component requires the service to avoid parsing safety refusals as valid JSON business records. Which approach passes this check?

  1. Reconcile batch responses by custom_id
  2. Supply previous turns on a new API request
  3. Handle a refusal stop_reason safely
  4. Call the Messages API with explicit model and max_tokens

Explanation

Correct Answer

C:

A refusal is a deliberate non-completion path that requires a defined application response; this behavior directly supports the need to avoid parsing safety refusals as valid JSON business records. Repeatedly increasing the token limit does not resolve a content refusal.

Incorrect Answers

A:

A caller-assigned identifier links each result to the original request independently of result ordering; the supported outcome is to match returned batch results to the right source documents despite varying completion order, whereas this scenario requires the ability to avoid parsing safety refusals as valid JSON business records; assuming completion order matches submission order risks mixing customer data.

B:

This approach serves the separate need to preserve a customer's follow-up reference to an earlier case number because stateless requests need relevant history passed back in the messages sequence; it does not establish the outcome to avoid parsing safety refusals as valid JSON business records, and a client connection alone does not give the model memory of previous calls.

D:

For the alternate objective to return an answer to a short user request through a backend service, a message request identifies the target model, a message list, and an output-token ceiling; the feature here must avoid parsing safety refusals as valid JSON business records, and the API is not a stateful chat session unless prior turns are supplied.

Question 228

A developer is reviewing alternatives for an engineering productivity feature that must serve many tenants without mixing their tool outputs or prior messages. Which option is a defensible fit?

  1. Choose isolated sandbox execution for untrusted code
  2. Keep user-specific state out of a shared agent process
  3. Use a durable session store
  4. Add idempotency keys to side-effecting operations

Explanation

Correct Answer

B:

Per-request identity and session boundaries prevent cross-tenant context leakage; this behavior directly supports the need to serve many tenants without mixing their tool outputs or prior messages. Shared conversation histories can inadvertently expose unrelated customers' data.

Incorrect Answers

A:

For the alternate objective to let an agent run generated scripts without exposing production host files, a restricted execution environment limits filesystem, network, and resource privileges; the feature here must serve many tenants without mixing their tool outputs or prior messages, and running code in the main application process expands the blast radius of a bad tool call.

C:

When an application needs to recover agent state when a container restarts during a lengthy customer case, external persistence holds the conversation and workflow metadata needed to resume safely; the current capability must instead serve many tenants without mixing their tool outputs or prior messages, while process-local variables disappear with the worker.

D:

Application-generated operation identifiers allow the downstream service to recognize a repeated request; the supported outcome is to avoid charging a customer twice after a worker retries a timed-out payment action, whereas this scenario requires the ability to serve many tenants without mixing their tool outputs or prior messages; model-level instructions cannot guarantee exactly-once execution after a partial failure.

Question 229

In an incident response project, acceptance test A checks whether a component can choose the right specialised handling path from the request category. Acceptance test B checks whether a different component can answer routine account questions without the overhead of orchestrating multiple agents. Which plan satisfies both tests?

  1. First component: Use an explicit completion condition and step budget. Second component: Restrict each specialist's tool permissions.
  2. First component: Apply routing before task execution. Second component: Use an explicit completion condition and step budget.
  3. First component: Apply routing before task execution. Second component: Keep a single agent with a small tool set.
  4. First component: Restrict each specialist's tool permissions. Second component: Keep a single agent with a small tool set.

Explanation

Correct Answer

C:

The first proposal fits its assigned role because an inexpensive classification or rule-based router sends each request to the applicable predefined workflow, while the second proposal fits its assigned role because a single constrained executor simplifies state, authentication, logging, and latency where specialization adds little; their respective acceptance criteria are to choose the right specialised handling path from the request category and to answer routine account questions without the overhead of orchestrating multiple agents.

Incorrect Answers

A:

The first proposal serves the different purpose to bound the runtime of an agent that repeatedly searches without converging instead of the assigned first role because a maximum iteration or cost budget plus a clear success criterion stops runaway tool loops, while the second proposal serves the different purpose to allow a research worker to read records without granting its coordinator's write privileges instead of the assigned second role because scoped tool capability prevents a delegated role from using authority unrelated to its assignment; their respective acceptance criteria are to choose the right specialised handling path from the request category and to answer routine account questions without the overhead of orchestrating multiple agents.

B:

The first proposal fits its assigned role because an inexpensive classification or rule-based router sends each request to the applicable predefined workflow, while the second proposal serves the different purpose to bound the runtime of an agent that repeatedly searches without converging instead of the assigned second role because a maximum iteration or cost budget plus a clear success criterion stops runaway tool loops; their respective acceptance criteria are to choose the right specialised handling path from the request category and to answer routine account questions without the overhead of orchestrating multiple agents.

D:

The first proposal serves the different purpose to allow a research worker to read records without granting its coordinator's write privileges instead of the assigned first role because scoped tool capability prevents a delegated role from using authority unrelated to its assignment, while the second proposal fits its assigned role because a single constrained executor simplifies state, authentication, logging, and latency where specialization adds little; their respective acceptance criteria are to choose the right specialised handling path from the request category and to answer routine account questions without the overhead of orchestrating multiple agents.

Question 230

Two features of a coding automation system are under review. One feature must establish a discoverable set of build and test conventions for a new Claude Code workspace; the other must share coding conventions with every collaborator entering the same repository. Which plan matches the respective needs?

  1. First component: Use headless invocation for automation. Second component: Use a project-level CLAUDE.md file.
  2. First component: Run repository initialization to create project guidance. Second component: Use a project-level CLAUDE.md file.
  3. First component: Define a custom slash command workflow. Second component: Use headless invocation for automation.
  4. First component: Run repository initialization to create project guidance. Second component: Define a custom slash command workflow.

Explanation

Correct Answer

B:

The first proposal fits its assigned role because repository initialization can generate or prepare project instruction material that the team subsequently reviews, while the second proposal fits its assigned role because a checked-in memory file supplies persistent project context across relevant coding sessions; their respective acceptance criteria are to establish a discoverable set of build and test conventions for a new Claude Code workspace and to share coding conventions with every collaborator entering the same repository.

Incorrect Answers

A:

The first proposal serves the different purpose to run Claude Code inside a CI job and capture its output without interactive prompts instead of the assigned first role because noninteractive execution lets scripts integrate agent work into existing pipelines with controlled formats, while the second proposal fits its assigned role because a checked-in memory file supplies persistent project context across relevant coding sessions; their respective acceptance criteria are to establish a discoverable set of build and test conventions for a new Claude Code workspace and to share coding conventions with every collaborator entering the same repository.

C:

The first proposal serves the different purpose to provide a consistent shortcut for a common repository maintenance task instead of the assigned first role because command-based workflows supply repeatable user-invoked instructions and arguments, while the second proposal serves the different purpose to run Claude Code inside a CI job and capture its output without interactive prompts instead of the assigned second role because noninteractive execution lets scripts integrate agent work into existing pipelines with controlled formats; their respective acceptance criteria are to establish a discoverable set of build and test conventions for a new Claude Code workspace and to share coding conventions with every collaborator entering the same repository.

D:

The first proposal fits its assigned role because repository initialization can generate or prepare project instruction material that the team subsequently reviews, while the second proposal serves the different purpose to provide a consistent shortcut for a common repository maintenance task instead of the assigned second role because command-based workflows supply repeatable user-invoked instructions and arguments; their respective acceptance criteria are to establish a discoverable set of build and test conventions for a new Claude Code workspace and to share coding conventions with every collaborator entering the same repository.

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