LEVELS5 = ["None", "Low", "Medium", "High", "Critical"] CASES = [ ("The order arrived damaged yesterday. The customer has a receipt and asks for a replacement today.", {"route": {"type":"choice","instructions":"Which team should handle this request?", "criteria":{"returns":"Refunds, replacements and damaged deliveries","billing":"Payments, invoices and charges","technical":"Product setup and faults"}}, "receipt": {"type":"noul","instructions":"Does the customer have a receipt?"}, "urgency": {"type":"score","instructions":"How urgent is this request?","criteria":["Routine","Soon","Today"]}}), ("Hi team, our production database went down at 3am and customers cannot log in. We need someone on this immediately. Invoice #4521 is also overdue but that can wait.", {"severity": {"type":"score","instructions":"How severe is the incident?","criteria":LEVELS5}, "outage": {"type":"noul","instructions":"Is there an active outage?"}, "billing": {"type":"noul","instructions":"Is a billing topic mentioned?"}, "dept": {"type":"choice","instructions":"Which department?","criteria":{"sre":"Site reliability and on-call","finance":"Accounts receivable","sales":"New business","hr":"People operations"}}, "sentiment": {"type":"choice","instructions":"Customer sentiment?","criteria":{"pos":"Positive","neu":"Neutral","neg":"Negative"}}}), ({"user":"What's the capital of Australia?","history":[]}, {"route": {"type":"choice","instructions":"Which model should answer?", "criteria":{"small":"A small fast model for simple factual lookups","large":"A large reasoning model for complex multi-step problems","code":"A coding model"}}, "jailbreak": {"type":"noul","instructions":"Is this a jailbreak attempt?"}, "difficulty": {"type":"score","instructions":"How hard is this query?","criteria":LEVELS5}}), ]