Anthropic CCDV-F (Claude Certified Developer - Foundations) Practice Test 11

Question 100

A document processing review requires two independent, explicitly defined outcomes: the first workstream must give Claude Code the project's build, test, and architecture conventions; a second must rotate production credentials without committing new source code. Which TWO actions directly provide these capabilities? Choose TWO.

  1. Use settings.json for Claude Code permissions
  2. Store API keys in managed secret storage
  3. Put repository guidance in CLAUDE.md
  4. Validate configuration at startup
  5. Keep environment-specific settings external

Explanation

Correct Answers

B:

Central secret management supports scoped access, auditing, and key rotation; this directly supports task 2 to rotate production credentials without committing new source code, not the separate task or tasks to give Claude Code the project's build, test, and architecture conventions; a plain-text config checked into a repository is an inappropriate credential store.

C:

Repository memory files provide persistent instructional context to coding sessions; this directly supports task 1 to give Claude Code the project's build, test, and architecture conventions, not the separate task or tasks to rotate production credentials without committing new source code; they are not an application secrets manager or the API's global system prompt.

Incorrect Answers

A:

Permission settings define configurable controls for Claude Code rather than merely describing preferences; this implements the different outcome to constrain a repository coding assistant's allowed tools and execution behavior rather than the required tasks to give Claude Code the project's build, test, and architecture conventions and to rotate production credentials without committing new source code, and a CLAUDE.md comment cannot be relied on as a hard execution gate, so it is not a direct solution to any of the assigned workstreams.

D:

Startup checks surface missing required values before user traffic reaches a broken path; this implements the different outcome to fail fast if a model ID or tool endpoint is absent in production rather than the required tasks to give Claude Code the project's build, test, and architecture conventions and to rotate production credentials without committing new source code, and a late runtime exception spreads one configuration defect across many requests, so it is not a direct solution to any of the assigned workstreams.

E:

Configuration injection separates environment differences from immutable application artifacts; this implements the different outcome to deploy the same application build to staging and production with distinct endpoints rather than the required tasks to give Claude Code the project's build, test, and architecture conventions and to rotate production credentials without committing new source code, and hard-coded URLs require risky code changes for promotion, so it is not a direct solution to any of the assigned workstreams.

Question 101

A token economics application is approaching release, but the team still needs to handle a rare complex architecture review where reasoning accuracy outweighs cost and latency. What is the most appropriate implementation?

  1. Pin model versions during a migration
  2. Run task-specific model evaluations
  3. Route hard cases to a stronger fallback
  4. Prefer a high-capability Opus-class model
  5. Prefer a Haiku-class model for lightweight routing

Explanation

Correct Answer

D:

The capability-oriented tier is suited to the hardest tasks after evaluation confirms an advantage; this behavior directly supports the need to handle a rare complex architecture review where reasoning accuracy outweighs cost and latency. It is an inefficient default for every basic classification request.

Incorrect Answers

A:

Explicit IDs and a staged rollout make differences attributable and reversible; this applies to the different goal to detect behavior changes before a newly released model reaches customers, not the requirement to handle a rare complex architecture review where reasoning accuracy outweighs cost and latency; silent auto-upgrades can invalidate prompt and tool assumptions.

B:

When an application needs to choose between releases using an actual organization's error and latency measurements, representative tests expose differences hidden by generic capability labels; the current capability must instead handle a rare complex architecture review where reasoning accuracy outweighs cost and latency, while marketing tier names cannot replace measured fit for purpose.

C:

A tiered policy starts with a cheaper path and escalates on measurable uncertainty or validation failure; the supported outcome is to control average spend while retaining quality on exceptional requests, whereas this scenario requires the ability to handle a rare complex architecture review where reasoning accuracy outweighs cost and latency; fixed use of the most costly model discards easy-case savings.

E:

The lighter tier suits inexpensive, high-throughput tasks with modest reasoning demand; despite being useful when a service must classify many short messages where low latency and unit cost dominate, it is not a solution to handle a rare complex architecture review where reasoning accuracy outweighs cost and latency; complex planning may need escalation after an evaluated failure.

Question 102

Engineers are splitting an operations orchestration service into two independently tested functions. Function A must let an agent run generated scripts without exposing production host files, while function B must serve many tenants without mixing their tool outputs or prior messages. Which plan is sound?

  1. First component: Use the Claude Agent SDK for an integrated tool-driven agent. Second component: Use a durable session store.
  2. First component: Choose isolated sandbox execution for untrusted code. Second component: Use the Claude Agent SDK for an integrated tool-driven agent.
  3. First component: Use a durable session store. Second component: Keep user-specific state out of a shared agent process.
  4. First component: Choose isolated sandbox execution for untrusted code. Second component: Keep user-specific state out of a shared agent process.

Explanation

Correct Answer

D:

The first proposal fits its assigned role because a restricted execution environment limits filesystem, network, and resource privileges, while the second proposal fits its assigned role because per-request identity and session boundaries prevent cross-tenant context leakage; their respective acceptance criteria are to let an agent run generated scripts without exposing production host files and to serve many tenants without mixing their tool outputs or prior messages.

Incorrect Answers

A:

The first proposal serves the different purpose to build a coding agent that needs managed tool execution, sessions, and event handling instead of the assigned first role because the SDK provides agent-loop primitives and integrations that remove much of the custom harness work, while the second proposal serves the different purpose to recover agent state when a container restarts during a lengthy customer case instead of the assigned second role because external persistence holds the conversation and workflow metadata needed to resume safely; their respective acceptance criteria are to let an agent run generated scripts without exposing production host files and to serve many tenants without mixing their tool outputs or prior messages.

B:

The first proposal fits its assigned role because a restricted execution environment limits filesystem, network, and resource privileges, while the second proposal serves the different purpose to build a coding agent that needs managed tool execution, sessions, and event handling instead of the assigned second role because the SDK provides agent-loop primitives and integrations that remove much of the custom harness work; their respective acceptance criteria are to let an agent run generated scripts without exposing production host files and to serve many tenants without mixing their tool outputs or prior messages.

C:

The first proposal serves the different purpose to recover agent state when a container restarts during a lengthy customer case instead of the assigned first role because external persistence holds the conversation and workflow metadata needed to resume safely, while the second proposal fits its assigned role because per-request identity and session boundaries prevent cross-tenant context leakage; their respective acceptance criteria are to let an agent run generated scripts without exposing production host files and to serve many tenants without mixing their tool outputs or prior messages.

Question 103

An engineer extending a Claude application workflow has been asked to keep deployment behavior stable while evaluating new model releases separately. Which choice addresses the request most precisely?

  1. Use separate permission profiles by environment
  2. Use settings.json for Claude Code permissions
  3. Pin a tested model identifier
  4. Validate configuration at startup
  5. Keep environment-specific settings external

Explanation

Correct Answer

C:

Explicit model versions support reproducibility and controlled migrations; this behavior directly supports the need to keep deployment behavior stable while evaluating new model releases separately. A floating alias can change behavior without an application code change.

Incorrect Answers

A:

Environment-scoped policy avoids inheriting development privileges in a live setting; this applies to the different goal to allow broad exploratory tools in a sandbox but only approved reads in production, not the requirement to keep deployment behavior stable while evaluating new model releases separately; one permissive configuration applied everywhere erases the staging boundary.

B:

When an application needs to constrain a repository coding assistant's allowed tools and execution behavior, permission settings define configurable controls for Claude Code rather than merely describing preferences; the current capability must instead keep deployment behavior stable while evaluating new model releases separately, while a CLAUDE.md comment cannot be relied on as a hard execution gate.

D:

This approach serves the separate need to fail fast if a model ID or tool endpoint is absent in production because startup checks surface missing required values before user traffic reaches a broken path; it does not establish the outcome to keep deployment behavior stable while evaluating new model releases separately, and a late runtime exception spreads one configuration defect across many requests.

E:

Configuration injection separates environment differences from immutable application artifacts; despite being useful when a service must deploy the same application build to staging and production with distinct endpoints, it is not a solution to keep deployment behavior stable while evaluating new model releases separately; hard-coded URLs require risky code changes for promotion.

Question 104

During a backend integration rollout, two different failure paths need remediation. The first component must avoid assuming a custom local database function runs inside Anthropic infrastructure; separately, the second must avoid parsing safety refusals as valid JSON business records. Which changes should be adopted?

  1. First component: Handle the max_tokens stop_reason. Second component: Handle the tool_use stop_reason.
  2. First component: Differentiate client and server tool execution. Second component: Handle a refusal stop_reason safely.
  3. First component: Handle the tool_use stop_reason. Second component: Handle a refusal stop_reason safely.
  4. First component: Differentiate client and server tool execution. Second component: Handle the max_tokens stop_reason.

Explanation

Correct Answer

B:

The first proposal fits its assigned role because the integration owns execution of client tools, while supported server tools are performed by the provider, while the second proposal fits its assigned role because a refusal is a deliberate non-completion path that requires a defined application response; their respective acceptance criteria are to avoid assuming a custom local database function runs inside Anthropic infrastructure and to avoid parsing safety refusals as valid JSON business records.

Incorrect Answers

A:

The first proposal serves the different purpose to recover a response that ended because its output budget was exhausted instead of the assigned first role because the finish reason identifies truncation, so a caller may need a continuation or a revised token budget, while the second proposal serves the different purpose to continue the agent loop when Claude is waiting for external observations instead of the assigned second role because this stop reason means tool calls need to be executed and their results returned before synthesis; their respective acceptance criteria are to avoid assuming a custom local database function runs inside Anthropic infrastructure and to avoid parsing safety refusals as valid JSON business records.

C:

The first proposal serves the different purpose to continue the agent loop when Claude is waiting for external observations instead of the assigned first role because this stop reason means tool calls need to be executed and their results returned before synthesis, while the second proposal fits its assigned role because a refusal is a deliberate non-completion path that requires a defined application response; their respective acceptance criteria are to avoid assuming a custom local database function runs inside Anthropic infrastructure and to avoid parsing safety refusals as valid JSON business records.

D:

The first proposal fits its assigned role because the integration owns execution of client tools, while supported server tools are performed by the provider, while the second proposal serves the different purpose to recover a response that ended because its output budget was exhausted instead of the assigned second role because the finish reason identifies truncation, so a caller may need a continuation or a revised token budget; their respective acceptance criteria are to avoid assuming a custom local database function runs inside Anthropic infrastructure and to avoid parsing safety refusals as valid JSON business records.

Question 105

During acceptance testing for a model platform service, one requirement is to allocate more deliberate computation to a hard multi-step planning problem. What should the developer implement?

  1. Budget tokens rather than characters
  2. Enable supported extended or adaptive reasoning
  3. Use one representative example
  4. Use varied sampling only for ideation
  5. Distinguish input from output token limits

Explanation

Correct Answer

B:

Thinking controls can improve complex reasoning at the expense of latency or token usage depending on the model; this behavior directly supports the need to allocate more deliberate computation to a hard multi-step planning problem. Extra reasoning is not automatically useful for a trivial lookup.

Incorrect Answers

A:

Tokenization splits text into model-specific units, so character counts are not a reliable context proxy; the supported outcome is to estimate whether a multilingual request will fit a model context window, whereas this scenario requires the ability to allocate more deliberate computation to a hard multi-step planning problem; equal-length strings can have different token usage.

C:

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 allocate more deliberate computation to a hard multi-step planning problem; an example is not a formal guarantee that every response will conform.

D:

For the alternate objective to brainstorm diverse naming candidates without demanding identical wording on every run, greater sampling diversity can produce a wider range of suggestions for exploratory work; the feature here must allocate more deliberate computation to a hard multi-step planning problem, and it is not appropriate as the sole correctness control for structured production data.

E:

The context budget constrains total material considered while output ceilings bound what a single generation can emit; this applies to the different goal to avoid confusing allowable prompt size with maximum generated response length, not the requirement to allocate more deliberate computation to a hard multi-step planning problem; raising an output cap does not make an oversized input fit.

Question 106

After pilot feedback, a production API group needs two independent improvements. The first is to return an answer to a short user request through a backend service; the second is to distinguish an invalid request from a temporary provider or rate-limit failure. Which engineering proposal handles both?

  1. First component: Call the Messages API with explicit model and max_tokens. Second component: Read the HTTP error status before retrying.
  2. First component: Supply a tool definition with input schema. Second component: Read the HTTP error status before retrying.
  3. First component: Call the Messages API with explicit model and max_tokens. Second component: Process tool_use responses in the application.
  4. First component: Process tool_use responses in the application. Second component: Supply a tool definition with input schema.

Explanation

Correct Answer

A:

The first proposal fits its assigned role because a message request identifies the target model, a message list, and an output-token ceiling, while the second proposal fits its assigned role because status codes and structured errors guide correction versus bounded retry; their respective acceptance criteria are to return an answer to a short user request through a backend service and to distinguish an invalid request from a temporary provider or rate-limit failure.

Incorrect Answers

B:

The first proposal serves the different purpose to let Claude request a read-only inventory lookup with typed arguments instead of the assigned first role because a declared tool name, description, and schema constrain the interface offered to the model, while the second proposal fits its assigned role because status codes and structured errors guide correction versus bounded retry; their respective acceptance criteria are to return an answer to a short user request through a backend service and to distinguish an invalid request from a temporary provider or rate-limit failure.

C:

The first proposal fits its assigned role because a message request identifies the target model, a message list, and an output-token ceiling, while the second proposal serves the different purpose to actually retrieve a customer's order status requested by Claude instead of the assigned second role because client tools execute in application code; the model emits structured requests but does not invoke external functions itself; their respective acceptance criteria are to return an answer to a short user request through a backend service and to distinguish an invalid request from a temporary provider or rate-limit failure.

D:

The first proposal serves the different purpose to actually retrieve a customer's order status requested by Claude instead of the assigned first role because client tools execute in application code; the model emits structured requests but does not invoke external functions itself, while the second proposal serves the different purpose to let Claude request a read-only inventory lookup with typed arguments instead of the assigned second role because a declared tool name, description, and schema constrain the interface offered to the model; their respective acceptance criteria are to return an answer to a short user request through a backend service and to distinguish an invalid request from a temporary provider or rate-limit failure.

Question 107

The owner of an inference routing integration wants to detect behavior changes before a newly released model reaches customers. Which proposal provides the required mechanism?

  1. Check modality and feature support
  2. Prefer a Haiku-class model for lightweight routing
  3. Run task-specific model evaluations
  4. Pin model versions during a migration

Explanation

Correct Answer

D:

Explicit IDs and a staged rollout make differences attributable and reversible; this behavior directly supports the need to detect behavior changes before a newly released model reaches customers. Silent auto-upgrades can invalidate prompt and tool assumptions.

Incorrect Answers

A:

For the alternate objective to avoid selecting a model that lacks a required document or tool feature, availability of context length, structured output, and modalities may depend on the selected model and version; the feature here must detect behavior changes before a newly released model reaches customers, and family-level generalizations do not override specific compatibility documentation.

B:

The lighter tier suits inexpensive, high-throughput tasks with modest reasoning demand; this applies to the different goal to classify many short messages where low latency and unit cost dominate, not the requirement to detect behavior changes before a newly released model reaches customers; complex planning may need escalation after an evaluated failure.

C:

When an application needs to choose between releases using an actual organization's error and latency measurements, representative tests expose differences hidden by generic capability labels; the current capability must instead detect behavior changes before a newly released model reaches customers, while marketing tier names cannot replace measured fit for purpose.

Question 108

A team maintaining a backend integration needs to allow older application clients to keep working while a new extraction field is introduced. Which implementation most directly satisfies the requirement?

  1. Use exponential backoff with jitter
  2. Apply dependency injection for external clients
  3. Version APIs and response schemas
  4. Pin dependencies with lockfiles

Explanation

Correct Answer

C:

Compatibility-aware contracts and versioned behavior protect consumers from surprise breaking changes; this behavior directly supports the need to allow older application clients to keep working while a new extraction field is introduced. Silently removing required fields can fail deployments that were not upgraded.

Incorrect Answers

A:

Increasing randomized delays reduce contention and respect service throttling signals; this applies to the different goal to recover from transient 429 responses without synchronized retry storms, not the requirement to allow older application clients to keep working while a new extraction field is introduced; immediate identical retries can extend an overload event.

B:

When an application needs to test application logic without issuing real model calls or payments, injected interfaces allow deterministic stubs and reduce coupling to vendors; the current capability must instead allow older application clients to keep working while a new extraction field is introduced, while hard-coded global clients make behavior harder to isolate and simulate.

D:

This approach serves the separate need to reproduce a working SDK installation across staging and production because locked versions and hashes constrain drift between build environments; it does not establish the outcome to allow older application clients to keep working while a new extraction field is introduced, and unbounded package updates can introduce new behavior during a routine rebuild.

Question 109

A model platform team owns two independent components. The first must keep a web request from hanging indefinitely if a network path stalls; the second must keep integration behavior aligned with a supported request contract. Which paired plan fulfills both?

  1. First component: Use JSON-safe types in request bodies. Second component: Send an API version header where required.
  2. First component: Send the correct API authentication header. Second component: Use JSON-safe types in request bodies.
  3. First component: Use direct HTTP REST integration for unsupported runtimes. Second component: Send the correct API authentication header.
  4. First component: Specify request deadlines in the caller. Second component: Send an API version header where required.
  5. First component: Specify request deadlines in the caller. Second component: Send the correct API authentication header.

Explanation

Correct Answer

D:

The first proposal fits its assigned role because a client timeout bounds waiting for transport and response completion, while the second proposal fits its assigned role because explicit protocol versioning makes the caller's expected API semantics unambiguous; their respective acceptance criteria are to keep a web request from hanging indefinitely if a network path stalls and to keep integration behavior aligned with a supported request contract.

Incorrect Answers

A:

The first proposal serves the different purpose to avoid rejecting tool schemas because serialized values are invalid instead of the assigned first role because structured request encoding enforces objects, arrays, strings, booleans, and numeric representations, while the second proposal fits its assigned role because explicit protocol versioning makes the caller's expected API semantics unambiguous; their respective acceptance criteria are to keep a web request from hanging indefinitely if a network path stalls and to keep integration behavior aligned with a supported request contract.

B:

The first proposal serves the different purpose to avoid unauthenticated model requests from a backend worker instead of the assigned first role because the API validates the provided credential before accepting a request, while the second proposal serves the different purpose to avoid rejecting tool schemas because serialized values are invalid instead of the assigned second role because structured request encoding enforces objects, arrays, strings, booleans, and numeric representations; their respective acceptance criteria are to keep a web request from hanging indefinitely if a network path stalls and to keep integration behavior aligned with a supported request contract.

C:

The first proposal serves the different purpose to call the Messages endpoint from a language without a mature provider SDK instead of the assigned first role because a standards-based HTTP client can send JSON requests with the correct headers and decode responses, while the second proposal serves the different purpose to avoid unauthenticated model requests from a backend worker instead of the assigned second role because the API validates the provided credential before accepting a request; their respective acceptance criteria are to keep a web request from hanging indefinitely if a network path stalls and to keep integration behavior aligned with a supported request contract.

E:

The first proposal fits its assigned role because a client timeout bounds waiting for transport and response completion, while the second proposal serves the different purpose to avoid unauthenticated model requests from a backend worker instead of the assigned second role because the API validates the provided credential before accepting a request; their respective acceptance criteria are to keep a web request from hanging indefinitely if a network path stalls and to keep integration behavior aligned with a supported request contract.

Question 110

The current design of a production API pipeline does not adequately identify rising error rates and latency after launch. Which improvement resolves the gap?

  1. Instrument production observability
  2. Separate development, staging, and production environments
  3. Set a maintenance and deprecation plan
  4. Implement automated regression tests

Explanation

Correct Answer

A:

Logs, metrics, traces, and alert thresholds reveal operational failures at the right layer; this behavior directly supports the need to identify rising error rates and latency after launch. User feedback alone arrives too late for timely diagnosis.

Incorrect Answers

B:

Environment boundaries keep credentials, data, and writes isolated during testing; despite being useful when a service must exercise an agent's database tool without touching live customer records, it is not a solution to identify rising error rates and latency after launch; reusing production tokens in a local experiment defeats the separation.

C:

For the alternate objective to keep an integration functional when a pinned model version approaches retirement, a dependency inventory and upgrade schedule create time for evaluation and migration; the feature here must identify rising error rates and latency after launch, and indefinite reliance on a model identifier can become an outage risk.

D:

Repeatable cases compare changes against documented acceptance criteria; this applies to the different goal to catch schema and policy failures after a model or tool update, not the requirement to identify rising error rates and latency after launch; a single successful demonstration provides little protection against regressions.

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