{"id":"cx-escalation-note-booking-repro","audience":"customer-support","category":"support-work","technique":"grounding","track":"support","level":"advanced","title":"Write a bug escalation with repro steps and a clean environment","free":false,"task":"Write an escalation for engineering with these labeled lines: Summary:, Steps:, Expected:, Actual:, Environment:, Impact:. Steps are exactly 3 numbered lines. For Environment use what the agent reproduced, not what the customer guessed. Leave out customer names and emails. Plain text, under 120 words.","materialLabel":"Your material (attached to your prompt automatically)","material":"TICKET #4410: online booking widget double-books\n\nCustomer (Priscilla D., priscilla@sunnysideyoga.example) wrote: \"Two people got the same 9am slot on Tuesday. I think it's because I'm on Chrome 118 on my old laptop. Fix it please. It happens when two people click Book almost together.\"\n\nAgent repro (Nov 3, Dev sandbox, Chrome 124, widget v3.2.1): Opened the same class slot in two browser windows, clicked Book in both within 1 second. Both bookings accepted; slot capacity is 1. Booking one at a time worked correctly (second was rejected).\n\nPattern: 14 customers have reported double-bookings in the last 3 days. All use widget v3.2.1, released Oct 30. Customers on v3.1 are fine.\n\nTemp workaround: none yet. Customers can manually cancel one booking.","rules":[{"type":"plainText"},{"type":"maxWords","max":120}],"mistakes":[{"title":"Using the customer's guess","why":"The customer blamed Chrome 118 and the AI takes that as the environment.","fix":"Say: \"Environment = the agent's repro, not the customer's guess.\"","checks":["r1"]},{"title":"Leaving out the scale","why":"The AI describes one booking and misses 14 customers and the v3.2.1 link.","fix":"Ask: \"Impact: how many customers, how long, which version.\"","checks":["r2"]},{"title":"Free-form story","why":"The AI retells the ticket in paragraphs.","fix":"Name the six labels and make Steps exactly 3 numbered lines.","checks":["r0"]},{"title":"Getting markdown back","why":"AI adds bold and headings by default, which show up as stray symbols where this is pasted.","fix":"Say \"plain text, no bold, no headings, no asterisks\".","checks":["plainText"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting and usually runs long, so a vague \"keep it short\" lands well over the limit.","fix":"Cap each part instead of the total, for example \"each line under 15 words\".","checks":["maxWords"]}],"promptHints":[{"label":"Asks for plain text","patterns":["plain","no bold","no markdown","no formatting","no headings"]},{"label":"Sets a length limit","patterns":["\\d+\\s*words","word (limit|count)","under \\d+"]},{"label":"Names the labels","patterns":["summary:","steps:","expected","environment"]},{"label":"Chooses the repro environment","patterns":["repro","agent","not the customer"]}],"hint":"Engineers need a clean repro. Name the labels, say which source to trust for the environment, and ask for the scale.","rubricLabels":["The output has the labeled lines Summary:, Steps:, Expected:, Actual:, Environment:, Impact: and Steps is exactly 3 numbered lines","Environment says Chrome 124 and widget v3.2.1 (the agent's repro) and does not cite Chrome 118 or the customer's old laptop","Impact says 14 customers in 3 days and that v3.1 is unaffected, and the output has no customer name or email"]}