Question 250 A user writes “refund invoice INV-731,” but the assistant searches all invoices because it never extracts the invoice identifier. Which verification is strongest for this particular failure mode? Score factuality and harmful-language behavior separately on adversarial tone prompts and confirm safe alternatives pass. Use two same-name fixture accounts and prove the agent abstains until an authorized unique account is selected. Test valid, malformed and missing IDs and inspect the resulting SQL parameters and permissions decisions. Test identical phrases in two timezones and verify no booking happens for ambiguous conversions. ExplanationCorrect Answer C:Entity extraction provides a stable lookup key; searching by broad natural language can leak or misselect records. Other Answers A:Test valid, malformed and missing IDs and inspect the resulting SQL parameters and permissions decisions. B:Test valid, malformed and missing IDs and inspect the resulting SQL parameters and permissions decisions. D:Test valid, malformed and missing IDs and inspect the resulting SQL parameters and permissions decisions.
Question 251 A user writes “refund invoice INV-731,” but the assistant searches all invoices because it never extracts the invoice identifier. What change to engineering governance would reduce this risk over time? Keep timezone-aware parsing and explicit confirmation rules at all booking side-effect boundaries. Standardize entity schemas and negative validation tests for record-specific tools. Maintain independent safety evaluation thresholds rather than collapsing every outcome into one overall correctness score. Require unique-ID resolution and entitlement filtering before personal financial retrieval. ExplanationCorrect Answer B:Entity extraction provides a stable lookup key; searching by broad natural language can leak or misselect records. Other Answers A:Entity extraction provides a stable lookup key; searching by broad natural language can leak or misselect records. C:This control addresses a different exposure: A customer-service assistant gives a helpful policy summary but uses harassing language toward the requester in its final paragraph. It would not prevent this failure, because entity extraction provides a stable lookup key; searching by broad natural language can leak or misselect records. D:That entity extraction provides a stable lookup key; searching by broad natural language can leak or misselect records.
Question 252 A travel agent hears “tomorrow morning in Tokyo” but calls a booking tool without resolving the user’s timezone or calendar date. Which engineering action most directly addresses this specific requirement? Require use of supplied specification fields, forbid unsupported values and include an explicit not_found response for missing attributes. Resolve the entity against authorized account identifiers and ask for disambiguation rather than choosing a similar name. Parse invoice_id as a typed field, validate its format and restrict the authorized lookup to that exact invoice. Extract date, time and location with timezone context, confirm ambiguous values, then invoke the availability tool. ExplanationCorrect Answer D:The action targets the stated cause: Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. Check the repair by: test identical phrases in two timezones and verify no booking happens for ambiguous conversions. Other Answers A:Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. B:Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. C:Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions.
Question 253 A travel agent hears “tomorrow morning in Tokyo” but calls a booking tool without resolving the user’s timezone or calendar date. What architectural constraint is most directly responsible? Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. General prompts allow parametric guesses to compete with retrieved evidence; instruction hierarchy must make source-grounded facts decisive. Entity extraction provides a stable lookup key; searching by broad natural language can leak or misselect records. Free-form names are not stable unique keys and can refer to multiple records with different access controls. ExplanationCorrect Answer A:Extract date, time and location with timezone context, confirm ambiguous values, then invoke the availability tool. Other Answers B:Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. C:That mechanism describes another problem: A user writes “refund invoice INV-731,” but the assistant searches all invoices because it never extracts the invoice identifier. It does not explain this case, where relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. D:Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions.
Question 254 A travel agent hears “tomorrow morning in Tokyo” but calls a booking tool without resolving the user’s timezone or calendar date. What validation evidence would most specifically confirm the resolution? Test valid, malformed and missing IDs and inspect the resulting SQL parameters and permissions decisions. Test identical phrases in two timezones and verify no booking happens for ambiguous conversions. Use two same-name fixture accounts and prove the agent abstains until an authorized unique account is selected. Test deliberately absent specifications and verify the answer says unavailable instead of supplying invented measurements. ExplanationCorrect Answer B:Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. Other Answers A:Test identical phrases in two timezones and verify no booking happens for ambiguous conversions. C:Test identical phrases in two timezones and verify no booking happens for ambiguous conversions. D:Test identical phrases in two timezones and verify no booking happens for ambiguous conversions.
Question 255 A travel agent hears “tomorrow morning in Tokyo” but calls a booking tool without resolving the user’s timezone or calendar date. What prevention measure is most targeted to this underlying cause? Keep timezone-aware parsing and explicit confirmation rules at all booking side-effect boundaries. Require unique-ID resolution and entitlement filtering before personal financial retrieval. Measure groundedness on missing-field and conflicting-source examples during every prompt revision. Standardize entity schemas and negative validation tests for record-specific tools. ExplanationCorrect Answer A:Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. Other Answers B:This control addresses a different exposure: A customer asks about the balance for “the Smith account,” and an agent inserts the first matching person into a finance prompt. It would not prevent this failure, because relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. C:That relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. D:Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions.
Question 256 A customer asks about the balance for “the Smith account,” and an agent inserts the first matching person into a finance prompt. Which implementation decision best restores the intended behavior? Parse invoice_id as a typed field, validate its format and restrict the authorized lookup to that exact invoice. Resolve the entity against authorized account identifiers and ask for disambiguation rather than choosing a similar name. Register the model or agent artifact and configuration revision with MLflow and link deployed versions to their evaluation runs. Extract date, time and location with timezone context, confirm ambiguous values, then invoke the availability tool. ExplanationCorrect Answer B:The action targets the stated cause: Free-form names are not stable unique keys and can refer to multiple records with different access controls. Check the repair by: use two same-name fixture accounts and prove the agent abstains until an authorized unique account is selected. Other Answers A:Free-form names are not stable unique keys and can refer to multiple records with different access controls. C:Free-form names are not stable unique keys and can refer to multiple records with different access controls. D:Free-form names are not stable unique keys and can refer to multiple records with different access controls.
Question 257 A customer asks about the balance for “the Smith account,” and an agent inserts the first matching person into a finance prompt. Which underlying technical principle explains why this occurs? Relative temporal language cannot be safely converted to a timestamp without locale and reference-date assumptions. Free-form names are not stable unique keys and can refer to multiple records with different access controls. Reproducible agent behavior depends on code, model, prompts and dependencies, not just the endpoint name. Entity extraction provides a stable lookup key; searching by broad natural language can leak or misselect records. ExplanationCorrect Answer B:Resolve the entity against authorized account identifiers and ask for disambiguation rather than choosing a similar name. Other Answers A:That mechanism describes another problem: A travel agent hears “tomorrow morning in Tokyo” but calls a booking tool without resolving the user’s timezone or calendar date. It does not explain this case, where free-form names are not stable unique keys and can refer to multiple records with different access controls. C:Free-form names are not stable unique keys and can refer to multiple records with different access controls. D:Free-form names are not stable unique keys and can refer to multiple records with different access controls.
Question 258 A customer asks about the balance for “the Smith account,” and an agent inserts the first matching person into a finance prompt. After correction, what targeted test should demonstrate success? Test identical phrases in two timezones and verify no booking happens for ambiguous conversions. Use two same-name fixture accounts and prove the agent abstains until an authorized unique account is selected. Test valid, malformed and missing IDs and inspect the resulting SQL parameters and permissions decisions. Restore the previous registered version in staging and verify trace metadata, dependencies and known-good acceptance tests. ExplanationCorrect Answer B:Free-form names are not stable unique keys and can refer to multiple records with different access controls. Other Answers A:Use two same-name fixture accounts and prove the agent abstains until an authorized unique account is selected. C:Use two same-name fixture accounts and prove the agent abstains until an authorized unique account is selected. D:Use two same-name fixture accounts and prove the agent abstains until an authorized unique account is selected.
Question 259 A customer asks about the balance for “the Smith account,” and an agent inserts the first matching person into a finance prompt. Which release safeguard most directly addresses this failure class? Standardize entity schemas and negative validation tests for record-specific tools. Require unique-ID resolution and entitlement filtering before personal financial retrieval. Keep timezone-aware parsing and explicit confirmation rules at all booking side-effect boundaries. Treat model and prompt artifact versioning as deployment inputs under change control. ExplanationCorrect Answer B:Free-form names are not stable unique keys and can refer to multiple records with different access controls. Other Answers A:That free-form names are not stable unique keys and can refer to multiple records with different access controls. C:Free-form names are not stable unique keys and can refer to multiple records with different access controls. D:This control addresses a different exposure: A project promotes agent code without recording dependency or prompt revision identifiers, making a rollback to known behavior impossible. It would not prevent this failure, because free-form names are not stable unique keys and can refer to multiple records with different access controls.
Question 260 A multilingual sales assistant extracts a product name but ignores the user's requested currency when preparing a quotation. Which design change directly satisfies the stated constraints? Extract date, time and location with timezone context, confirm ambiguous values, then invoke the availability tool. Parse invoice_id as a typed field, validate its format and restrict the authorized lookup to that exact invoice. Require use of supplied specification fields, forbid unsupported values and include an explicit not_found response for missing attributes. Separate product_id, locale and requested_currency fields, fetch the approved price and convert only through an auditable exchange-rate service. ExplanationCorrect Answer D:The action targets the stated cause: Named entities and user preferences must map to different validated parameters; generation should not invent price conversions. Check the repair by: replay two currencies and verify quoted amounts reconcile with the authorized rate and product source. Other Answers A:Named entities and user preferences must map to different validated parameters; generation should not invent price conversions. B:Named entities and user preferences must map to different validated parameters; generation should not invent price conversions. C:Named entities and user preferences must map to different validated parameters; generation should not invent price conversions.
