Anthropic CCDV-F (Claude Certified Developer - Foundations) Practice Test 67

Question 660

A Claude application delivery milestone depends on two separate capabilities. One module must activate a new tool execution path for only internal testers; a second module must serve independent API calls without blocking a worker thread for each network wait. Which implementation option is appropriate?

  1. First component: Use feature flags for risky behavior changes. Second component: Use asynchronous I/O for concurrent requests.
  2. First component: Use feature flags for risky behavior changes. Second component: Refactor duplicated request orchestration.
  3. First component: Use pull requests and focused code review. Second component: Use asynchronous I/O for concurrent requests.
  4. First component: Refactor duplicated request orchestration. Second component: Use pull requests and focused code review.

Explanation

Correct Answer

A:

The first proposal fits its assigned role because runtime controls can segment rollout and permit quick disabling without a code rebuild, while the second proposal fits its assigned role because async execution overlaps I/O-bound tasks while maintaining control over concurrency; their respective acceptance criteria are to activate a new tool execution path for only internal testers and to serve independent API calls without blocking a worker thread for each network wait.

Incorrect Answers

B:

The first proposal fits its assigned role because runtime controls can segment rollout and permit quick disabling without a code rebuild, while the second proposal serves the different purpose to maintain multiple Claude entry points without drifting retry and logging behavior instead of the assigned second role because a shared, tested client layer centralizes cross-cutting concerns and reduces inconsistent fixes; their respective acceptance criteria are to activate a new tool execution path for only internal testers and to serve independent API calls without blocking a worker thread for each network wait.

C:

The first proposal serves the different purpose to detect unsafe permissions and schema changes before merging an agent tool instead of the assigned first role because peer review creates an explicit checkpoint for correctness, security, and maintainability, while the second proposal fits its assigned role because async execution overlaps I/O-bound tasks while maintaining control over concurrency; their respective acceptance criteria are to activate a new tool execution path for only internal testers and to serve independent API calls without blocking a worker thread for each network wait.

D:

The first proposal serves the different purpose to maintain multiple Claude entry points without drifting retry and logging behavior instead of the assigned first role because a shared, tested client layer centralizes cross-cutting concerns and reduces inconsistent fixes, while the second proposal serves the different purpose to detect unsafe permissions and schema changes before merging an agent tool instead of the assigned second role because peer review creates an explicit checkpoint for correctness, security, and maintainability; their respective acceptance criteria are to activate a new tool execution path for only internal testers and to serve independent API calls without blocking a worker thread for each network wait.

Question 661

After pilot feedback, a document processing group needs two independent improvements. The first is to prevent an untested tool-allowlist edit from reaching all developers immediately; the second is to fail fast if a model ID or tool endpoint is absent in production. Which engineering proposal handles both?

  1. First component: Store API keys in managed secret storage. Second component: Validate configuration at startup.
  2. First component: Track MCP server versions and contracts. Second component: Store API keys in managed secret storage.
  3. First component: Roll out configuration through reviewed changes. Second component: Validate configuration at startup.
  4. First component: Roll out configuration through reviewed changes. Second component: Track MCP server versions and contracts.

Explanation

Correct Answer

C:

The first proposal fits its assigned role because change review and versioning provide accountability and a tested rollback path, while the second proposal fits its assigned role because startup checks surface missing required values before user traffic reaches a broken path; their respective acceptance criteria are to prevent an untested tool-allowlist edit from reaching all developers immediately and to fail fast if a model ID or tool endpoint is absent in production.

Incorrect Answers

A:

The first proposal serves the different purpose to rotate production credentials without committing new source code instead of the assigned first role because central secret management supports scoped access, auditing, and key rotation, while the second proposal fits its assigned role because startup checks surface missing required values before user traffic reaches a broken path; their respective acceptance criteria are to prevent an untested tool-allowlist edit from reaching all developers immediately and to fail fast if a model ID or tool endpoint is absent in production.

B:

The first proposal serves the different purpose to keep external tool connectors compatible after client upgrades instead of the assigned first role because declared versions and integration tests expose resource or schema changes before deployment, while the second proposal serves the different purpose to rotate production credentials without committing new source code instead of the assigned second role because central secret management supports scoped access, auditing, and key rotation; their respective acceptance criteria are to prevent an untested tool-allowlist edit from reaching all developers immediately and to fail fast if a model ID or tool endpoint is absent in production.

D:

The first proposal fits its assigned role because change review and versioning provide accountability and a tested rollback path, while the second proposal serves the different purpose to keep external tool connectors compatible after client upgrades instead of the assigned second role because declared versions and integration tests expose resource or schema changes before deployment; their respective acceptance criteria are to prevent an untested tool-allowlist edit from reaching all developers immediately and to fail fast if a model ID or tool endpoint is absent in production.

Question 662

Two separate tickets in an engineering productivity backlog require different controls. One needs to branch between remediation and escalation based on the result of a security scan; the other needs to select orchestration dependencies for a project that already has a stable task engine. Which implementation pair is correct?

  1. First component: Use a graph with conditional edges. Second component: Use a framework only when its primitives fit.
  2. First component: Use a graph with conditional edges. Second component: Use event logs for tool actions.
  3. First component: Use event logs for tool actions. Second component: Use a graceful fallback path.
  4. First component: Use retrieval before generation. Second component: Use event logs for tool actions.
  5. First component: Use a graceful fallback path. Second component: Use a framework only when its primitives fit.

Explanation

Correct Answer

A:

The first proposal fits its assigned role because conditional graph transitions encode known branches while allowing model reasoning within selected nodes, while the second proposal fits its assigned role because frameworks can help with graphs, retries, and checkpoints but create maintenance and integration costs; their respective acceptance criteria are to branch between remediation and escalation based on the result of a security scan and to select orchestration dependencies for a project that already has a stable task engine.

Incorrect Answers

B:

The first proposal fits its assigned role because conditional graph transitions encode known branches while allowing model reasoning within selected nodes, while the second proposal serves the different purpose to explain why an agent modified a configuration during a failed automated release instead of the assigned second role because structured traces connect tool intent, arguments, results, and subsequent decisions for audit and replay; their respective acceptance criteria are to branch between remediation and escalation based on the result of a security scan and to select orchestration dependencies for a project that already has a stable task engine.

C:

The first proposal serves the different purpose to explain why an agent modified a configuration during a failed automated release instead of the assigned first role because structured traces connect tool intent, arguments, results, and subsequent decisions for audit and replay, while the second proposal serves the different purpose to keep a customer service application operating when an expensive specialist agent cannot complete instead of the assigned second role because a bounded failure branch can return a safe partial result or route to a human rather than loop indefinitely; their respective acceptance criteria are to branch between remediation and escalation based on the result of a security scan and to select orchestration dependencies for a project that already has a stable task engine.

D:

The first proposal serves the different purpose to answer questions from changing internal documentation with traceable source excerpts instead of the assigned first role because retrieval grounds a model answer in current authoritative text rather than hoping parametric knowledge is current, while the second proposal serves the different purpose to explain why an agent modified a configuration during a failed automated release instead of the assigned second role because structured traces connect tool intent, arguments, results, and subsequent decisions for audit and replay; their respective acceptance criteria are to branch between remediation and escalation based on the result of a security scan and to select orchestration dependencies for a project that already has a stable task engine.

E:

The first proposal serves the different purpose to keep a customer service application operating when an expensive specialist agent cannot complete instead of the assigned first role because a bounded failure branch can return a safe partial result or route to a human rather than loop indefinitely, while the second proposal fits its assigned role because frameworks can help with graphs, retries, and checkpoints but create maintenance and integration costs; their respective acceptance criteria are to branch between remediation and escalation based on the result of a security scan and to select orchestration dependencies for a project that already has a stable task engine.

Question 663

A developer is reviewing alternatives for a model platform feature that must estimate whether a multilingual request will fit a model context window. Which option is a defensible fit?

  1. Budget tokens rather than characters
  2. Treat model outputs as stochastic
  3. Use low-temperature generation for repeatable extraction
  4. Use varied sampling only for ideation
  5. Use one representative example

Explanation

Correct Answer

A:

Tokenization splits text into model-specific units, so character counts are not a reliable context proxy; this behavior directly supports the need to estimate whether a multilingual request will fit a model context window. Equal-length strings can have different token usage.

Incorrect Answers

B:

When an application needs to design tests that tolerate benign wording variation while detecting semantic failures, non-determinism requires evaluation by criteria and repeated examples rather than exact-string equality everywhere; the current capability must instead estimate whether a multilingual request will fit a model context window, while a single run is insufficient to prove robust model behavior.

C:

Lower randomness often improves consistency for narrow, constrained tasks; the supported outcome is to reduce variation when a service maps documents into consistent categories, whereas this scenario requires the ability to estimate whether a multilingual request will fit a model context window; it does not guarantee identical output or replace schema validation.

D:

This approach serves the separate need to brainstorm diverse naming candidates without demanding identical wording on every run because greater sampling diversity can produce a wider range of suggestions for exploratory work; it does not establish the outcome to estimate whether a multilingual request will fit a model context window, and it is not appropriate as the sole correctness control for structured production data.

E:

A single demonstration can anchor the expected shape and tone of a generated response; despite being useful when a service must clarify a specific formatting convention that a plain instruction was not conveying, it is not a solution to estimate whether a multilingual request will fit a model context window; an example is not a formal guarantee that every response will conform.

Question 664

A design review identifies a missing capability in an output validation project: it must preserve auditability when an extractor reformats a date or currency. Which change meets that goal?

  1. Validate required schema fields
  2. Escape generated content in the UI
  3. Keep raw and normalized values distinct
  4. Handle refusal and empty content explicitly

Explanation

Correct Answer

C:

Storing the original evidence and validated normalized value supports later reconciliation; this behavior directly supports the need to preserve auditability when an extractor reformats a date or currency. Overwriting the raw source eliminates a useful verification trail.

Incorrect Answers

A:

For the alternate objective to reject a response that omits a mandated account identifier, schema checks enforce field presence, types, and enumerations before the result enters a database; the feature here must preserve auditability when an extractor reformats a date or currency, and syntactically valid JSON may still be contract-invalid.

B:

Output encoding or sanitization protects display contexts from untrusted markup; this applies to the different goal to show model-produced HTML snippets without allowing browser script injection, not the requirement to preserve auditability when an extractor reformats a date or currency; a correct JSON wrapper does not make embedded HTML safe.

D:

Response handling must branch on stop reason and content block types before decoding data; the supported outcome is to prevent a parser crash when the model returns no ordinary text, whereas this scenario requires the ability to preserve auditability when an extractor reformats a date or currency; assuming every call yields a usable business record creates hidden error paths.

Question 665

Two separate tickets in an agent governance backlog require different controls. One needs to avoid committing provider credentials inside a shared repository; the other needs to cut off an exposed token discovered in CI logs. Which implementation pair is correct?

  1. First component: Separate human and service identities. Second component: Authorize by requested resource and action.
  2. First component: Load keys from a managed secret service. Second component: Separate human and service identities.
  3. First component: Load keys from a managed secret service. Second component: Rotate a compromised API key promptly.
  4. First component: Authorize by requested resource and action. Second component: Rotate a compromised API key promptly.

Explanation

Correct Answer

C:

The first proposal fits its assigned role because secrets managers support controlled retrieval, rotation, and access logs, while the second proposal fits its assigned role because revocation plus replacement ends use of the leaked credential and requires checking for misuse; their respective acceptance criteria are to avoid committing provider credentials inside a shared repository and to cut off an exposed token discovered in CI logs.

Incorrect Answers

A:

The first proposal serves the different purpose to attribute tool actions to the right workload rather than a developer's personal account instead of the assigned first role because distinct principals simplify least privilege, revocation, and audit attribution, while the second proposal serves the different purpose to deny access to a different customer's records even when the caller is authenticated instead of the assigned second role because authorization checks must evaluate scope for each operation, not only whether a login exists; their respective acceptance criteria are to avoid committing provider credentials inside a shared repository and to cut off an exposed token discovered in CI logs.

B:

The first proposal fits its assigned role because secrets managers support controlled retrieval, rotation, and access logs, while the second proposal serves the different purpose to attribute tool actions to the right workload rather than a developer's personal account instead of the assigned second role because distinct principals simplify least privilege, revocation, and audit attribution; their respective acceptance criteria are to avoid committing provider credentials inside a shared repository and to cut off an exposed token discovered in CI logs.

D:

The first proposal serves the different purpose to deny access to a different customer's records even when the caller is authenticated instead of the assigned first role because authorization checks must evaluate scope for each operation, not only whether a login exists, while the second proposal fits its assigned role because revocation plus replacement ends use of the leaked credential and requires checking for misuse; their respective acceptance criteria are to avoid committing provider credentials inside a shared repository and to cut off an exposed token discovered in CI logs.

Question 666

An engineer extending an agent platform workflow has been asked to investigate an open-ended incident where the next query depends on earlier findings. Which choice addresses the request most precisely?

  1. Use sequential chaining for dependent subtasks
  2. Use parallel fan-out and a reconciliation step
  3. Use an agentic decision loop
  4. Use a central state record for orchestration

Explanation

Correct Answer

C:

A model-directed loop can choose subsequent actions based on tool observations until it reaches a stopping condition; this behavior directly supports the need to investigate an open-ended incident where the next query depends on earlier findings. A predetermined chain cannot reliably anticipate all branches of an exploratory diagnosis.

Incorrect Answers

A:

When an application needs to ensure a generated specification is validated before code generation begins, each stage consumes an approved artifact from the previous stage, preserving dependency order and stage-specific checks; the current capability must instead investigate an open-ended incident where the next query depends on earlier findings, while parallel execution would allow downstream work to proceed on unverified assumptions.

B:

Parallel tasks reduce critical-path latency when they do not depend on one another; a join step resolves conflicting findings; the supported outcome is to analyze several independent document sections concurrently and combine the results, whereas this scenario requires the ability to investigate an open-ended incident where the next query depends on earlier findings; a serial chain makes independent stages wait without improving correctness.

D:

Persisted job state tracks completed stages, pending actions, and recoverable checkpoints outside the model context; despite being useful when a service must resume a multi-step agentic process after worker or network failure, it is not a solution to investigate an open-ended incident where the next query depends on earlier findings; relying solely on transient conversation messages risks duplicated side effects after restart.

Question 667

An engineering productivity review requires two independent, explicitly defined outcomes: the first workstream must continue a support investigation after many rounds of log inspection; a second must detect mistakes in an agent-generated SQL migration before running it. Which TWO actions directly provide these capabilities? Choose TWO.

  1. Use event logs for tool actions
  2. Use a graceful fallback path
  3. Summarize long-running state with checkpoints
  4. Use a plan-execute-review loop
  5. Use an independent verifier step

Explanation

Correct Answers

C:

Periodic compaction preserves decisions, open questions, and evidence references within the context budget; this directly supports task 1 to continue a support investigation after many rounds of log inspection, not the separate task or tasks to detect mistakes in an agent-generated SQL migration before running it; blindly truncating oldest messages can remove critical constraints.

E:

A separate validation stage checks the artifact against explicit tests and constraints, limiting self-confirming errors; this directly supports task 2 to detect mistakes in an agent-generated SQL migration before running it, not the separate task or tasks to continue a support investigation after many rounds of log inspection; asking the original agent whether it is correct is weaker than executable verification.

Incorrect Answers

A:

Structured traces connect tool intent, arguments, results, and subsequent decisions for audit and replay; this implements the different outcome to explain why an agent modified a configuration during a failed automated release rather than the required tasks to detect mistakes in an agent-generated SQL migration before running it and to continue a support investigation after many rounds of log inspection, and a final chat summary omits the exact causal sequence, so it is not a direct solution to any of the assigned workstreams.

B:

A bounded failure branch can return a safe partial result or route to a human rather than loop indefinitely; this implements the different outcome to keep a customer service application operating when an expensive specialist agent cannot complete rather than the required tasks to detect mistakes in an agent-generated SQL migration before running it and to continue a support investigation after many rounds of log inspection, and unbounded retries do not ensure eventual success, so it is not a direct solution to any of the assigned workstreams.

D:

Separating plan, execution, and review makes deviations visible before a final outcome is accepted; this implements the different outcome to manage a complex migration that requires checkpoints after several tool actions rather than the required tasks to detect mistakes in an agent-generated SQL migration before running it and to continue a support investigation after many rounds of log inspection, and a one-shot response cannot verify environmental effects, so it is not a direct solution to any of the assigned workstreams.

Question 668

The owner of a customer platform integration wants to allow broad exploratory tools in a sandbox but only approved reads in production. Which proposal provides the required mechanism?

  1. Use separate permission profiles by environment
  2. Version prompt templates alongside code
  3. Validate configuration at startup
  4. Store API keys in managed secret storage
  5. Pin a tested model identifier

Explanation

Correct Answer

A:

Environment-scoped policy avoids inheriting development privileges in a live setting; this behavior directly supports the need to allow broad exploratory tools in a sandbox but only approved reads in production. One permissive configuration applied everywhere erases the staging boundary.

Incorrect Answers

B:

This approach serves the separate need to trace a quality regression to the exact instruction revision used because source-controlled prompt assets make rollbacks, diffs, and evaluations possible; it does not establish the outcome to allow broad exploratory tools in a sandbox but only approved reads in production, and editing prompts in an unversioned dashboard hides causal history.

C:

Startup checks surface missing required values before user traffic reaches a broken path; despite being useful when a service must fail fast if a model ID or tool endpoint is absent in production, it is not a solution to allow broad exploratory tools in a sandbox but only approved reads in production; a late runtime exception spreads one configuration defect across many requests.

D:

For the alternate objective to rotate production credentials without committing new source code, central secret management supports scoped access, auditing, and key rotation; the feature here must allow broad exploratory tools in a sandbox but only approved reads in production, and a plain-text config checked into a repository is an inappropriate credential store.

E:

Explicit model versions support reproducibility and controlled migrations; this applies to the different goal to keep deployment behavior stable while evaluating new model releases separately, not the requirement to allow broad exploratory tools in a sandbox but only approved reads in production; a floating alias can change behavior without an application code change.

Question 669

A team maintaining an incident response needs to bound the runtime of an agent that repeatedly searches without converging. Which implementation most directly satisfies the requirement?

  1. Use parallel fan-out and a reconciliation step
  2. Introduce a human approval checkpoint
  3. Use an explicit completion condition and step budget
  4. Use a central state record for orchestration

Explanation

Correct Answer

C:

A maximum iteration or cost budget plus a clear success criterion stops runaway tool loops; this behavior directly supports the need to bound the runtime of an agent that repeatedly searches without converging. Merely asking the model to finish quickly does not enforce execution limits.

Incorrect Answers

A:

This approach serves the separate need to analyze several independent document sections concurrently and combine the results because parallel tasks reduce critical-path latency when they do not depend on one another; a join step resolves conflicting findings; it does not establish the outcome to bound the runtime of an agent that repeatedly searches without converging, and a serial chain makes independent stages wait without improving correctness.

B:

A human gate separates analysis from externally consequential action and gives a reviewer a veto; despite being useful when a service must prevent an autonomous investigation from issuing irreversible account changes without review, it is not a solution to bound the runtime of an agent that repeatedly searches without converging; a system-prompt request to be careful is not an enforceable approval workflow.

D:

Persisted job state tracks completed stages, pending actions, and recoverable checkpoints outside the model context; this applies to the different goal to resume a multi-step agentic process after worker or network failure, not the requirement to bound the runtime of an agent that repeatedly searches without converging; relying solely on transient conversation messages risks duplicated side effects after restart.

Question 670

An incident response delivery milestone depends on two separate capabilities. One module must continue a support investigation after many rounds of log inspection; a second module must answer questions from changing internal documentation with traceable source excerpts. Which implementation option is appropriate?

  1. First component: Summarize long-running state with checkpoints. Second component: Delegate narrow subtasks with a separate context.
  2. First component: Delegate narrow subtasks with a separate context. Second component: Use a graph with conditional edges.
  3. First component: Use a graph with conditional edges. Second component: Use retrieval before generation.
  4. First component: Summarize long-running state with checkpoints. Second component: Use retrieval before generation.

Explanation

Correct Answer

D:

The first proposal fits its assigned role because periodic compaction preserves decisions, open questions, and evidence references within the context budget, while the second proposal fits its assigned role because retrieval grounds a model answer in current authoritative text rather than hoping parametric knowledge is current; their respective acceptance criteria are to continue a support investigation after many rounds of log inspection and to answer questions from changing internal documentation with traceable source excerpts.

Incorrect Answers

A:

The first proposal fits its assigned role because periodic compaction preserves decisions, open questions, and evidence references within the context budget, while the second proposal serves the different purpose to protect a coordinator's window from huge specialist tool outputs instead of the assigned second role because each subagent can work in a bounded context and return a concise, verified artifact; their respective acceptance criteria are to continue a support investigation after many rounds of log inspection and to answer questions from changing internal documentation with traceable source excerpts.

B:

The first proposal serves the different purpose to protect a coordinator's window from huge specialist tool outputs instead of the assigned first role because each subagent can work in a bounded context and return a concise, verified artifact, while the second proposal serves the different purpose to branch between remediation and escalation based on the result of a security scan instead of the assigned second role because conditional graph transitions encode known branches while allowing model reasoning within selected nodes; their respective acceptance criteria are to continue a support investigation after many rounds of log inspection and to answer questions from changing internal documentation with traceable source excerpts.

C:

The first proposal serves the different purpose to branch between remediation and escalation based on the result of a security scan instead of the assigned first role because conditional graph transitions encode known branches while allowing model reasoning within selected nodes, while the second proposal fits its assigned role because retrieval grounds a model answer in current authoritative text rather than hoping parametric knowledge is current; their respective acceptance criteria are to continue a support investigation after many rounds of log inspection and to answer questions from changing internal documentation with traceable source excerpts.

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