Question 950
For a multi-repository troubleshooting agent, a reviewer identifies two independent obligations: all contributors receive the repository's stable engineering conventions and the report is not drafted before evidence is checked. Which proposal should be approved?
- put team-wide rules only in one developer's user-level instructions; and place evidence validation before the drafting stage
- put team-wide rules only in one developer's user-level instructions; and start drafting at the same time as the source search and assume success
- put concise shared guidance in the repository root CLAUDE.md; and start drafting at the same time as the source search and assume success
- put concise shared guidance in the repository root CLAUDE.md; and place evidence validation before the drafting stage
- implement only put concise shared guidance in the repository root CLAUDE.md; make no corresponding change for the second requirement
Explanation
Correct Answer
D:In a multi-repository troubleshooting agent, project instructions travel with the codebase and establish common architecture and build constraints; alongside this, a dependent stage must receive validated inputs to avoid propagating unsupported conclusions, satisfying the distinct requirements that all contributors receive the repository's stable engineering conventions and that the report is not drafted before evidence is checked.
Incorrect Answers
A:In a multi-repository troubleshooting agent, a dependent stage must receive validated inputs to avoid propagating unsupported conclusions, but personal files do not distribute the intended conventions to other contributors; therefore the separate requirement that all contributors receive the repository's stable engineering conventions would remain unsatisfied.
B:In a multi-repository troubleshooting agent, personal files do not distribute the intended conventions to other contributors and the writer may treat unverified or absent evidence as real; the design would fail both conditions: all contributors receive the repository's stable engineering conventions and the report is not drafted before evidence is checked.
C:In a multi-repository troubleshooting agent, project instructions travel with the codebase and establish common architecture and build constraints, but the writer may treat unverified or absent evidence as real; therefore the separate requirement that the report is not drafted before evidence is checked would remain unsatisfied.
E:In a multi-repository troubleshooting agent, project instructions travel with the codebase and establish common architecture and build constraints, but the independent mechanism for the other criterion is absent: a dependent stage must receive validated inputs to avoid propagating unsupported conclusions; as written, the design does not ensure that the report is not drafted before evidence is checked.
Question 951
Tests for an order-cancellation service will pass only if a failed MCP operation is visible as an error to the client and if schema-valid output is also checked against source evidence. Which proposed design directly satisfies both checks?
- mark the tool result as an error and supply structured failure details; and assume a schema-conforming response is factually correct
- return a success-shaped value saying the operation probably worked; and validate semantic constraints and provenance in application logic
- return a success-shaped value saying the operation probably worked; and assume a schema-conforming response is factually correct
- mark the tool result as an error and supply structured failure details; and validate semantic constraints and provenance in application logic
Explanation
Correct Answer
D:In an order-cancellation service, explicit error status separates failed execution from valid data content; alongside this, JSON shape does not prove that an extracted claim is supported by the document, satisfying the distinct requirements that a failed MCP operation is visible as an error to the client and that schema-valid output is also checked against source evidence.
Incorrect Answers
A:In an order-cancellation service, explicit error status separates failed execution from valid data content, but syntactic conformance cannot guarantee source-level truth; therefore the separate requirement that schema-valid output is also checked against source evidence would remain unsatisfied.
B:In an order-cancellation service, JSON shape does not prove that an extracted claim is supported by the document, but downstream agents can mistake a failure for verified business data; therefore the separate requirement that a failed MCP operation is visible as an error to the client would remain unsatisfied.
C:In an order-cancellation service, downstream agents can mistake a failure for verified business data and syntactic conformance cannot guarantee source-level truth; the design would fail both conditions: a failed MCP operation is visible as an error to the client and schema-valid output is also checked against source evidence.
Question 952
For a merchant support assistant, acceptance depends on demonstrating that refund execution is blocked until customer identity is verified. Which approach is the strongest fit?
- enforce a prerequisite gate in application logic before the write tool
- use an idempotency key and confirm the prior commit before retrying
- ask the assistant in prose to remember verification first
- route the exception to a human approval step before execution
- send a structured state packet with completed checks and unresolved decisions
Explanation
Correct Answer
A:For a merchant support assistant, deterministic preconditions protect an irreversible operation even if the model proposes it early; this directly establishes that refund execution is blocked until customer identity is verified instead of relying on an unrelated workflow control.
Incorrect Answers
B:Although side-effecting operations need deduplication against timeouts and repeated messages, that only supports the distinct goal that repeated delivery of one action request cannot create duplicate writes, not the requirement that refund execution is blocked until customer identity is verified in a merchant support assistant.
C:In a merchant support assistant, a prompt is advisory and does not technically block an unauthorized tool call; as a result, the stated requirement that refund execution is blocked until customer identity is verified would still not be met.
D:Although approval is a workflow boundary separate from language-model confidence, that only supports the distinct goal that out-of-policy exceptions receive an authorized decision, not the requirement that refund execution is blocked until customer identity is verified in a merchant support assistant.
E:Although a typed transition preserves obligations and facts between workflow stages, that only supports the distinct goal that a specialist handoff retains its prerequisites and evidence, not the requirement that refund execution is blocked until customer identity is verified in a merchant support assistant.
Question 953
In a code policy conformance job, the immediate design requirement is that a consequential claim receives independent human sign-off. Which change most directly establishes that behavior?
- calibrate confidence against known outcomes and source support
- log the reviewer identity, decision, justification, and artifact version
- put uncertain cases in an explicit review queue with priority and context
- route high-impact or low-evidence results to an authorized reviewer
- accept model confidence as a substitute for authorization
Explanation
Correct Answer
D:For a code policy conformance job, human approval supplies accountability where a mistaken model result has material cost; this directly establishes that a consequential claim receives independent human sign-off instead of relying on an unrelated workflow control.
Incorrect Answers
A:Although observed error rates and evidence quality are better than a fluent self-rating, that only supports the distinct goal that confidence scores reflect available evidence and uncertainty, not the requirement that a consequential claim receives independent human sign-off in a code policy conformance job.
B:Although a versioned approval trail supports diagnosis and later governance reviews, that only supports the distinct goal that reviewer changes remain auditable, not the requirement that a consequential claim receives independent human sign-off in a code policy conformance job.
C:Although triage preserves unattended processing for low-risk cases while catching exceptions, that only supports the distinct goal that an ambiguous item is separated from routine automation, not the requirement that a consequential claim receives independent human sign-off in a code policy conformance job.
E:In a code policy conformance job, a model score does not confer business or legal authority; as a result, the stated requirement that a consequential claim receives independent human sign-off would still not be met.
Question 954
When designing a warranty claims assistant, the requirement is not satisfied unless tool responses are associated with the exact invocations that produced them and every high-impact tool invocation produces an audit event. Which implementation addresses the two independent conditions?
- return tool_result blocks using the original tool_use_id values; and record the tool name, permitted inputs, outcome, and task identifier in a hook
- return tool results with newly invented identifiers; and infer tool activity from the final natural-language answer
- return tool results with newly invented identifiers; and record the tool name, permitted inputs, outcome, and task identifier in a hook
- return tool_result blocks using the original tool_use_id values; and infer tool activity from the final natural-language answer
Explanation
Correct Answer
A:In a warranty claims assistant, correlation identifiers bind asynchronous or parallel results to the right requests; alongside this, runtime instrumentation observes actual proposed or completed tool activity, satisfying the distinct requirements that tool responses are associated with the exact invocations that produced them and that every high-impact tool invocation produces an audit event.
Incorrect Answers
B:In a warranty claims assistant, the model cannot reliably associate a result with an outstanding call and a summary can omit failed calls or delegated actions; the design would fail both conditions: tool responses are associated with the exact invocations that produced them and every high-impact tool invocation produces an audit event.
C:In a warranty claims assistant, runtime instrumentation observes actual proposed or completed tool activity, but the model cannot reliably associate a result with an outstanding call; therefore the separate requirement that tool responses are associated with the exact invocations that produced them would remain unsatisfied.
D:In a warranty claims assistant, correlation identifiers bind asynchronous or parallel results to the right requests, but a summary can omit failed calls or delegated actions; therefore the separate requirement that every high-impact tool invocation produces an audit event would remain unsatisfied.
Question 955
The architect for a legacy service decomposition needs to ensure that developers can explicitly invoke an established project procedure. What should the team implement?
- write undocumented one-off ad hoc prompts for each engineer
- declare allowed-tools for the skill and enforce executor permissions
- make the skill available under a recognizable slash invocation
- create a documented reusable skill in SKILL.md
Explanation
Correct Answer
C:For a legacy service decomposition, a discoverable workflow gives the team a consistent entry point for repetitive work; this directly establishes that developers can explicitly invoke an established project procedure instead of relying on an unrelated workflow control.
Incorrect Answers
A:In a legacy service decomposition, unshared prompts make results hard to reproduce and review; as a result, the stated requirement that developers can explicitly invoke an established project procedure would still not be met.
B:Although tool restrictions narrow the agent's operational capability while the skill runs, that only supports the distinct goal that a high-risk workflow only invokes specifically approved capabilities, not the requirement that developers can explicitly invoke an established project procedure in a legacy service decomposition.
D:Although skills package instructions and supporting resources for invocation across similar tasks, that only supports the distinct goal that a repeatable code-review procedure is available as a named workflow, not the requirement that developers can explicitly invoke an established project procedure in a legacy service decomposition.
Question 956
An audit of a large-codebase architecture explorer identifies two unmet requirements: a broad tool catalog does not overwhelm a small task and one developer keeps an MCP server private to their machine. Select TWO actions that remedy those particular deficiencies.
- commit private workstation commands into the repository's shared MCP file
- supply only context-relevant tools for the current agent
- use a user-local MCP configuration rather than committing project configuration
- load all available MCP tools into every specialist turn
Explanation
Correct Answers
B:For a large-codebase architecture explorer, reducing irrelevant tool definitions improves selection and saves context, which directly establishes that a broad tool catalog does not overwhelm a small task; the separate requirement that one developer keeps an MCP server private to their machine is handled by a different action.
C:For a large-codebase architecture explorer, user scope is appropriate for personal endpoints and avoids distributing local paths, which directly establishes that one developer keeps an MCP server private to their machine; the separate requirement that a broad tool catalog does not overwhelm a small task is handled by a different action.
Incorrect Answers
A:For a large-codebase architecture explorer, machine-specific commands can break teammates' environments, so this option cannot establish that one developer keeps an MCP server private to their machine; correctly addressing a broad tool catalog does not overwhelm a small task elsewhere does not repair that deficiency.
D:For a large-codebase architecture explorer, tool bloat makes the available actions harder to discriminate, so this option cannot establish that a broad tool catalog does not overwhelm a small task; correctly addressing one developer keeps an MCP server private to their machine elsewhere does not repair that deficiency.
Question 957
A team operating a multi-source competitor research system wants to establish two separate behaviors: the system can continue when one specialist returns an explicit failure; in addition, specialized agents work with only relevant permissions. Which plan implements both?
- record per-branch status and let the coordinator choose retry or partial response; and define each subagent role and expose the minimum matching tools
- treat every subagent error string as a valid research conclusion; and define each subagent role and expose the minimum matching tools
- record per-branch status and let the coordinator choose retry or partial response; and give every subagent the coordinator's complete tool inventory
- treat every subagent error string as a valid research conclusion; and give every subagent the coordinator's complete tool inventory
Explanation
Correct Answer
A:In a multi-source competitor research system, the orchestration layer must distinguish missing evidence from a negative factual finding; alongside this, tool scoping shrinks both decision confusion and the blast radius of delegated actions, satisfying the distinct requirements that the system can continue when one specialist returns an explicit failure and that specialized agents work with only relevant permissions.
Incorrect Answers
B:In a multi-source competitor research system, tool scoping shrinks both decision confusion and the blast radius of delegated actions, but error text is operational state rather than evidence about the world; therefore the separate requirement that the system can continue when one specialist returns an explicit failure would remain unsatisfied.
C:In a multi-source competitor research system, the orchestration layer must distinguish missing evidence from a negative factual finding, but unneeded write or network tools expand risk without improving the specialist task; therefore the separate requirement that specialized agents work with only relevant permissions would remain unsatisfied.
D:In a multi-source competitor research system, error text is operational state rather than evidence about the world and unneeded write or network tools expand risk without improving the specialist task; the design would fail both conditions: the system can continue when one specialist returns an explicit failure and specialized agents work with only relevant permissions.
Question 958
Before expanding a dependency upgrade initiative, the owner must demonstrate that every high-impact tool invocation produces an audit event, while also demonstrating that a large investigation spans unrelated subsystems. Which design provides both controls?
- record the tool name, permitted inputs, outcome, and task identifier in a hook; and delegate bounded subsystem searches and merge path-based findings
- record the tool name, permitted inputs, outcome, and task identifier in a hook; and give every specialist the complete monorepo transcript
- implement only record the tool name, permitted inputs, outcome, and task identifier in a hook; make no corresponding change for the second requirement
- infer tool activity from the final natural-language answer; and give every specialist the complete monorepo transcript
- infer tool activity from the final natural-language answer; and delegate bounded subsystem searches and merge path-based findings
Explanation
Correct Answer
A:In a dependency upgrade initiative, runtime instrumentation observes actual proposed or completed tool activity; alongside this, subsystem partitioning permits parallel work without duplicating full repository context, satisfying the distinct requirements that every high-impact tool invocation produces an audit event and that a large investigation spans unrelated subsystems.
Incorrect Answers
B:In a dependency upgrade initiative, runtime instrumentation observes actual proposed or completed tool activity, but each branch pays the same large context cost and increases noise; therefore the separate requirement that a large investigation spans unrelated subsystems would remain unsatisfied.
C:In a dependency upgrade initiative, runtime instrumentation observes actual proposed or completed tool activity, but the independent mechanism for the other criterion is absent: subsystem partitioning permits parallel work without duplicating full repository context; as written, the design does not ensure that a large investigation spans unrelated subsystems.
D:In a dependency upgrade initiative, a summary can omit failed calls or delegated actions and each branch pays the same large context cost and increases noise; the design would fail both conditions: every high-impact tool invocation produces an audit event and a large investigation spans unrelated subsystems.
E:In a dependency upgrade initiative, subsystem partitioning permits parallel work without duplicating full repository context, but a summary can omit failed calls or delegated actions; therefore the separate requirement that every high-impact tool invocation produces an audit event would remain unsatisfied.
Question 959
Production readiness for a change-impact analysis service depends on ensuring that independent stages of a complex job run efficiently and that an existing file receives a precise localized replacement. Select the most appropriate combined approach.
- divide work along independent outputs and run those stages in parallel; and overwrite the whole file with a newly generated copy
- divide work along independent outputs and run those stages in parallel; and use Edit for an exact targeted change
- split the job into an arbitrary fixed chain of agents; and overwrite the whole file with a newly generated copy
- split the job into an arbitrary fixed chain of agents; and use Edit for an exact targeted change
Explanation
Correct Answer
B:In a change-impact analysis service, dependency-aware decomposition avoids serial work when branches do not consume one another; alongside this, surgical edits preserve unrelated lines and provide reviewable diffs, satisfying the distinct requirements that independent stages of a complex job run efficiently and that an existing file receives a precise localized replacement.
Incorrect Answers
A:In a change-impact analysis service, dependency-aware decomposition avoids serial work when branches do not consume one another, but full replacement risks losing unrelated code and formatting; therefore the separate requirement that an existing file receives a precise localized replacement would remain unsatisfied.
C:In a change-impact analysis service, a rigid sequence creates waits even where no dependency exists and full replacement risks losing unrelated code and formatting; the design would fail both conditions: independent stages of a complex job run efficiently and an existing file receives a precise localized replacement.
D:In a change-impact analysis service, surgical edits preserve unrelated lines and provide reviewable diffs, but a rigid sequence creates waits even where no dependency exists; therefore the separate requirement that independent stages of a complex job run efficiently would remain unsatisfied.
Question 960
For a repository change-review job, a reviewer identifies two independent obligations: sensitive tool calls are denied before they reach an executor and workload growth does not cause uncontrolled parallel fan-out. Which proposal should be approved?
- write a stronger instruction in the system prompt; and launch all possible child tasks at once without resource controls
- register a PreToolUse hook that checks the operation against policy; and limit concurrency and schedule independent jobs against execution budgets
- write a stronger instruction in the system prompt; and limit concurrency and schedule independent jobs against execution budgets
- register a PreToolUse hook that checks the operation against policy; and launch all possible child tasks at once without resource controls
Explanation
Correct Answer
B:In a repository change-review job, pre-execution interception can reject a disallowed call rather than relying on post-hoc advice; alongside this, bounded orchestration balances throughput with tool quotas and downstream service capacity, satisfying the distinct requirements that sensitive tool calls are denied before they reach an executor and that workload growth does not cause uncontrolled parallel fan-out.
Incorrect Answers
A:In a repository change-review job, an instruction alone does not provide deterministic execution-time authorization and unbounded fan-out can overwhelm tools and blow through cost budgets; the design would fail both conditions: sensitive tool calls are denied before they reach an executor and workload growth does not cause uncontrolled parallel fan-out.
C:In a repository change-review job, bounded orchestration balances throughput with tool quotas and downstream service capacity, but an instruction alone does not provide deterministic execution-time authorization; therefore the separate requirement that sensitive tool calls are denied before they reach an executor would remain unsatisfied.
D:In a repository change-review job, pre-execution interception can reject a disallowed call rather than relying on post-hoc advice, but unbounded fan-out can overwhelm tools and blow through cost budgets; therefore the separate requirement that workload growth does not cause uncontrolled parallel fan-out would remain unsatisfied.