Anthropic CCA-F (Claude Certified Architect - Foundations (CCAR-F)) Practice Test 15

Question 140

A design review automates multi-stage document and code checks. A novel repository may require more or fewer search steps. What should the team do to resolve this particular gap?

  1. Stop dependent deployment stages and report the failed security gate
  2. Launch checks concurrently while preserving distinct result records
  3. Use a join gate that requires successful results from both checks
  4. Use an explicit test checklist and deterministic execution order
  5. Use budgeted iterative retrieval with stopping conditions

Explanation

Correct Answer

E:

Unknown structure calls for conditional exploration, but it must not continue indefinitely after finding sufficient evidence.

Incorrect Answers

A:

A blocker should halt stages that cannot succeed safely; continuing merely consumes resources and may create risk; it protects against the case where a failed security check makes downstream deployment work irrelevant, rather than the failure described here: a novel repository may require more or fewer search steps.

B:

Checks that read the same immutable revision can reduce total time without introducing dependency errors; this is useful for independent lint and unit-test suites can run on an unchanged build, but it leaves the stated requirement to adapt exploration depth to observed complexity unmet.

C:

Approval before every prerequisite completes risks accepting a change that violates an unchecked requirement; that applies when a final reviewer needs both lint and tests to pass, whereas this case requires a response to a novel repository may require more or fewer search steps.

D:

This measure fits a different trigger: a fixed compliance check requires a predictable list of tests. In that setting, Stable compliance steps should be reproducible in code, with Claude assisting individual reasoning tasks rather than inventing the pipeline; it does not resolve the present need to adapt exploration depth to observed complexity.

Question 141

A development agent explores competing fixes across multiple sessions. Two proposed fixes need separate experiments from the same baseline. Which control is most directly relevant?

  1. Rerun relevant tests on the current revision before accepting success
  2. Resume the relevant session and inspect the recorded pending work
  3. Use separate worktrees or sandboxes for mutable branch experiments
  4. Fork a common session context for each experiment

Explanation

Correct Answer

D:

Branching keeps baseline assumptions comparable, whereas serially editing one conversation blends conclusions. Here it serves the explicit need to compare alternatives without shared prompt drift without adding a separate unverified inference.

Incorrect Answers

A:

Tests describe a specific code state; a prior green result cannot establish correctness for a changed repository; that applies when a resumed worker reports success based on a cached test from last week, whereas this case requires a response to two proposed fixes need separate experiments from the same baseline.

B:

This measure fits a different trigger: a developer needs to continue yesterday's unfinished refactor. In that setting, Continuing the actual conversation preserves intent and decisions, although the current code must still be validated; it does not resolve the present need to compare alternatives without shared prompt drift.

C:

Forked dialogue history alone does not isolate disk writes; workspace separation prevents cross-experiment contamination; this is useful for one experiment altered local files that another experiment must not see, but it leaves the stated requirement to compare alternatives without shared prompt drift unmet.

Question 142

A support assistant drafts responses that must meet measurable service standards. An agent answers a refund inquiry without checking customer eligibility. How should the workflow be adjusted?

  1. Specify required facts, tone limits, forbidden claims, and measurable completion checks
  2. Require explicit uncertainty and an escalation route when no sourced estimate exists
  3. Require policy lookup and verified eligibility fields before recommending a refund
  4. Specify acknowledgement, verified status, and a concrete authorized next action

Explanation

Correct Answer

C:

The response contract should identify the evidence needed for a decision instead of rewarding confident language unsupported by account data.

Incorrect Answers

A:

This measure fits a different trigger: reviewers disagree because the instruction merely says 'be helpful.'. In that setting, Explicit criteria guide generation and review; subjective adjectives alone do not define whether a reply satisfied the support policy; it does not resolve the present need to make evidence prerequisites explicit in the prompt.

B:

A safe prompt states what to do when evidence is absent; a request for confidence or completeness can otherwise encourage fabrication; that applies when the assistant invents a delivery estimate when the carrier has not supplied one, whereas this case requires a response to an agent answers a refund inquiry without checking customer eligibility.

D:

Component-level criteria make the outcome testable while leaving room for natural wording within the approved service tone; this is useful for a support reply must acknowledge an issue and give the next permitted step, but it leaves the stated requirement to make evidence prerequisites explicit in the prompt unmet.

Question 143

An application extracts structured fields from invoices with Claude. The deployment uses a Claude model and settings that support forcing a designated tool, and every extracted record must be submitted through that function. Which implementation decision is the best fit?

  1. Define an array of line-item objects with required amount and description fields
  2. Declare amount as a number and reject nonnumeric values
  3. On a model that supports it, set tool_choice to tool with the required function name
  4. Have Claude supply fields through a tool interface with a defined JSON Schema

Explanation

Correct Answer

C:

A forced tool choice can enforce this output channel on compatible models; the application must still validate arguments and should use schema-constrained output when forced tool use is unavailable.

Incorrect Answers

A:

Nested arrays preserve line-item boundaries and permit validation of each element without relying on free-text delimiters; that applies when the system stores several line items for one invoice, whereas this case requires a response to the deployment uses a Claude model and settings that support forcing a designated tool, and every extracted record must be submitted through that function.

B:

Type constraints prevent textual approximations from being silently interpreted as ledger-ready amounts; it protects against the case where a paid amount is emitted as a phrase such as 'about fifty', rather than the failure described here: the deployment uses a Claude model and settings that support forcing a designated tool, and every extracted record must be submitted through that function.

D:

This measure fits a different trigger: free-form JSON text frequently contains trailing commentary that breaks parsing. In that setting, Tool-argument schemas provide a clearer machine-readable contract than merely asking for 'JSON only' in prose, though validation remains necessary; it does not resolve the present need to constrain supported tool selection to a designated function.

Question 144

Claude Code is working with an unfamiliar repository too large to read at once. Search results span many modules but only two files are relevant to the current bug. What should the architect specify in the design?

  1. Record directories, ownership boundaries, and key entry points before deep reading
  2. Store the current hypothesis, inspected files, pending test, and next action
  3. Read the two implicated files and retain references to the other hits
  4. Retain constraints, decisions, evidence references, and remaining risks

Explanation

Correct Answer

C:

Focused context makes the debugging task tractable without deleting the ability to revisit excluded evidence. For the stated case, the check is whether the system can narrow the active working set; this approach supplies that control.

Incorrect Answers

A:

This measure fits a different trigger: an agent cannot tell where major services live in a large monorepo. In that setting, A file map supports targeted exploration and prevents repeatedly flooding context with unrelated source files; it does not resolve the present need to narrow the active working set.

B:

A resumable scratchpad prevents rediscovery and reduces the risk of treating an old hypothesis as a verified fix; that applies when a lengthy investigation is paused before the final patch, whereas this case requires a response to search results span many modules but only two files are relevant to the current bug.

D:

Useful compaction shortens history while keeping the information needed to continue safely and verify the eventual result; it protects against the case where the working context is compacted after many exploratory reads, rather than the failure described here: search results span many modules but only two files are relevant to the current bug.

Question 145

An architect is reviewing Claude Code memory usage for a large engineering organization. An instruction is mandatory only for the billing service. Which change most directly addresses this condition?

  1. Keep the billing convention in the billing project or scoped rule
  2. Store the durable personal convention in the user's Claude Code instruction scope
  3. Put the shared standard in the repository-level CLAUDE.md
  4. Reference the maintained build configuration rather than hard-code versions in memory
  5. Move genuinely team-wide requirements into version-controlled project guidance

Explanation

Correct Answer

A:

Narrow placement keeps tasks in unrelated services from inheriting incompatible commands or assumptions. The relevant requirement is to avoid applying service rules to other teams, and this measure operates at that decision point.

Incorrect Answers

B:

This measure fits a different trigger: an instruction is relevant to every repository maintained by one developer. In that setting, User guidance follows the developer across projects while remaining separate from team-wide repository policy; it does not resolve the present need to avoid applying service rules to other teams.

C:

A root project instruction communicates cross-cutting conventions without repeating them in every package; this is useful for a convention applies to every service in the same monorepo, but it leaves the stated requirement to avoid applying service rules to other teams unmet.

D:

The build manifest is authoritative for version details; repeating values in instructions creates stale conflicting copies; it protects against the case where a rule changes frequently because its tool version is pinned elsewhere, rather than the failure described here: an instruction is mandatory only for the billing service.

E:

Local-only user notes may not travel with the repository; team enforcement needs a shared reviewed artifact; that applies when a teammate edits private local guidance and expects it in code review, whereas this case requires a response to an instruction is mandatory only for the billing service.

Question 146

An enterprise support assistant uses lengthy conversation and policy histories. A chat history includes several superseded account addresses. What is the most defensible next step?

  1. Fetch the latest verified address and label old entries as historical
  2. Retrieve the pertinent excerpt and carry its location into the answer
  3. Treat retrieved text as evidence, not an override of higher-priority controls
  4. Maintain a versioned decision ledger with approval status and source-turn references

Explanation

Correct Answer

A:

Chronological conversation content can contain stale facts; an authoritative lookup prevents outdated messages from driving action.

Incorrect Answers

B:

Focused evidence helps the model judge a rule accurately and allows a reviewer to verify that the conclusion follows from the source; it protects against the case where an agent answers from a giant document without quoting the relevant passage, rather than the failure described here: a chat history includes several superseded account addresses.

C:

External material can be adversarial or incorrect; long context must not promote a source document into policy authority; that applies when a retrieved source conflicts with an earlier system-defined safety requirement, whereas this case requires a response to a chat history includes several superseded account addresses.

D:

A decision ledger tracks which resolution was authorized and which proposals were reversed, preventing a condensed conversation from treating tentative advice as binding; this is useful for the support chat contains several competing ticket resolutions, including proposals that were later rejected, but it leaves the stated requirement to identify current authoritative state unmet.

Question 147

A development team wants consistent Claude Code behavior across a shared monorepo. One developer prefers a personal response style that colleagues do not share. Which approach satisfies the stated constraint?

  1. Inspect project and user guidance together and resolve conflicting directions explicitly
  2. Place concise project instructions in a committed project CLAUDE.md file
  3. Keep the personal preference in user-level Claude Code guidance
  4. Keep a short index of essential rules and reference detailed docs only when needed

Explanation

Correct Answer

C:

User-level configuration supports individual habits without imposing personal formatting conventions on the entire repository.

Incorrect Answers

A:

Overlapping memory files may supply inconsistent directions; an architect should make shared requirements clear rather than hope the model infers priority; it protects against the case where the team changes a shared convention but contributors have conflicting personal notes, rather than the failure described here: one developer prefers a personal response style that colleagues do not share.

B:

This measure fits a different trigger: every contributor must follow the same build and review commands for the repository. In that setting, Project instructions make conventions available when working in the repository; they should describe relevant workflow rules rather than replicate all documentation; it does not resolve the present need to separate user preference from team policy.

D:

Large generic guidance competes with working context; focused instructions are easier for Claude to use reliably; that applies when the main CLAUDE.md duplicates a lengthy internal engineering handbook, whereas this case requires a response to one developer prefers a personal response style that colleagues do not share.

Question 148

An MCP client connects to business-data tools provided by third-party servers. A tool is marked read-only in descriptive metadata but its implementation writes audit records. Which mechanism should be selected?

  1. Treat tool annotations as hints and evaluate real side effects
  2. Enforce caller authorization in the client and backend regardless of tool descriptions
  3. Prevent retrieved text from authorizing tool calls or changing instruction hierarchy
  4. Mark pending operations unknown and reconnect or reconcile through supported state

Explanation

Correct Answer

A:

Capability annotations improve discovery but are not a security boundary or proof that code has no writes. Here it serves the explicit need to verify behavior instead of relying solely on annotations without adding a separate unverified inference.

Incorrect Answers

B:

This measure fits a different trigger: a server's advertised description says it may ignore the user's data-access restrictions. In that setting, Server metadata is untrusted discovery data; access checks must remain enforceable even if metadata is malicious or misleading; it does not resolve the present need to verify behavior instead of relying solely on annotations.

C:

Prompt-injection material may arrive through normal tool results, so the agent's higher-priority policies and allowlists remain in force; it protects against the case where an MCP server returns a copied webpage containing hostile 'system' instructions, rather than the failure described here: a tool is marked read-only in descriptive metadata but its implementation writes audit records.

D:

Transport failure does not establish that a side effect failed; the client must track request outcomes and avoid blind mutation retries; this is useful for a remote MCP server drops connection during a multi-step task, but it leaves the stated requirement to verify behavior instead of relying solely on annotations unmet.

Question 149

A customer-service agent uses hooks around tool operations. A source-search tool returns inconsistent date formats that confuse downstream parsing. What should be implemented at the appropriate boundary?

  1. Transform dates at a controlled post-tool boundary into a canonical format
  2. Inspect and reject paths outside the approved workspace in pre-tool logic
  3. Record the event at the tool-execution boundary using a hook or service log
  4. Remove or deny shell execution in the actual tool permissions or hook

Explanation

Correct Answer

A:

A post-tool normalization step makes structured evidence consistent without hoping a later prompt corrects every variant.

Incorrect Answers

B:

Argument-level checks prevent an otherwise permitted write tool from escaping the authorized directory; it protects against the case where a dangerous path appears in file-write arguments before a tool runs, rather than the failure described here: a source-search tool returns inconsistent date formats that confuse downstream parsing.

C:

Execution-time logging captures real calls and outcomes; model-generated explanations alone are not reliable audit evidence; this is useful for the application needs a durable record whenever a privileged tool is invoked, but it leaves the stated requirement to normalize tool output before model consumption unmet.

D:

Text guidance is probabilistic, whereas execution controls can reject a forbidden operation regardless of model output; that applies when a prompt says 'never run shell commands,' but the tool remains available and callable, whereas this case requires a response to a source-search tool returns inconsistent date formats that confuse downstream parsing.

Question 150

A CI pipeline runs Claude Code non-interactively for pull-request review. A downstream CI job must parse the review result reliably. Which option best meets the operational need?

  1. Run in a restricted sandbox with minimal tokens and no secret-bearing privileged steps
  2. Set job timeout, token or iteration budgets, and explicit failure behavior
  3. Use an independent check or human approval before merging
  4. Request a defined JSON output format and validate its schema in the pipeline
  5. Run Claude Code in noninteractive print mode with a bounded task prompt

Explanation

Correct Answer

D:

Structured output supports deterministic gating and avoids brittle parsing of free-form review prose. The relevant requirement is to emit structured machine-readable output, and this measure operates at that decision point.

Incorrect Answers

A:

Pull-request content is untrusted input; prompt instructions cannot substitute for runner isolation or secret scoping; this is useful for a pull request from an untrusted fork contains instructions to exfiltrate secrets, but it leaves the stated requirement to emit structured machine-readable output unmet.

B:

CI should fail visibly when budgets are exhausted rather than hang or silently return partial success; it protects against the case where a CI task might continue indefinitely under repeated tool calls, rather than the failure described here: a downstream CI job must parse the review result reliably.

C:

A model reviewing its own output is correlated with its mistakes; a distinct gate reduces unchecked self-approval; that applies when the team wants Claude's review to recommend fixes but not approve its own changes, whereas this case requires a response to a downstream CI job must parse the review result reliably.

E:

This measure fits a different trigger: the build runner has no interactive terminal for answering Claude prompts. In that setting, A headless execution path avoids UI expectations and lets CI capture the completed response as an artifact; it does not resolve the present need to emit structured machine-readable output.

Claude Certified Architect - Foundations (CCAR-F) (CCA-F) Practice Tests