{"id":"rec-scorecard-rating-evidence-gap-table","audience":"recruiting","category":"people-work","track":"people","level":"advanced","technique":"verification","title":"Flag scorecard ratings that have no written evidence","free":false,"task":"Check the interviewer's scorecard and return a table with the header Criterion | Rating | Evidence written | Flag. One row per criterion in the scorecard order. Evidence written is Yes if the notes contain a specific example or observation for that criterion, otherwise No. Flag is exactly \"OK\" if Evidence is Yes, or \"Needs evidence\" if No, or \"Off-topic\" if the notes mention anything about family, age, health, accent or appearance. Do not change any rating, rank, or recommend a decision. Plain text.","materialLabel":"One interviewer's scorecard and notes (attached to your prompt automatically)","material":"Candidate: Ibrahim Saleh. Interviewer: Fern Wallace. Role: Accounts Payable Specialist.\n\nRatings (1 to 4)\n- Accuracy and attention to detail: 4\n- Vendor communication: 3\n- Process improvement: 4\n- Teamwork: 3\n\nNotes\n- Accuracy: Walked through how he found a duplicate $2,300 payment during a month-end review and set up a weekly check.\n- Vendor communication: Said he is good with vendors.\n- Process improvement: Great energy.\n- Teamwork: He has a young family and says he will be flexible with the team. Described covering a colleague's close tasks in March.\n\nThe interviewer wrote the ratings before the debrief. The panel makes the decision.","rules":[{"type":"maxWords","max":170},{"type":"plainText"}],"mistakes":[{"title":"Accepting vague notes","why":"The AI treats 'good with vendors' as evidence.","fix":"Say: \"Evidence is Yes only for a specific example or observation.\"","checks":["r1"]},{"title":"Missing the family mention","why":"The AI marks teamwork OK because of the March example.","fix":"Say: \"Flag Off-topic if notes mention family, age, health, accent or appearance.\"","checks":["r2"]},{"title":"Changing ratings","why":"The AI suggests new scores or a decision.","fix":"Say: \"Do not change ratings or recommend.\"","checks":["r2"]},{"title":"Header","why":"The AI changes columns.","fix":"Say: \"Header exactly Criterion | Rating | Evidence written | Flag.\"","checks":["r0"]},{"title":"Trusting the AI to count words","why":"AI is bad at counting. Ask for a total word limit and it often lands over it.","fix":"Cap each part instead, for example: \"each line under 15 words.\"","checks":["maxWords"]},{"title":"Markdown","why":"The AI bolds words or adds headings.","fix":"Say: \"Plain text, no bold, no headings.\"","checks":["plainText"]}],"promptHints":[{"label":"Exact header","patterns":["criterion \\|","evidence written","flag"]},{"label":"Flag values","patterns":["ok","needs evidence","off-topic"]},{"label":"Specific evidence","patterns":["specific example"]},{"label":"Ratings unchanged","patterns":["do not change"]},{"label":"Word cap","patterns":["under \\d+ words"]}],"hint":"Give the header, the exact flag values and the test for evidence.","mission":"Flag scorecard ratings that lack written evidence.","musts":["Specific example needed","Flag off-topic notes","Don't change ratings"],"rubricLabels":["Has the header Criterion | Rating | Evidence written | Flag and exactly four rows in scorecard order with the interviewer's ratings unchanged (4, 3, 4, 3)","Evidence written and Flag are right: accuracy Yes / OK, vendor communication No / Needs evidence, process improvement No / Needs evidence","Teamwork is flagged Off-topic because the notes mention his family, and the output makes no change to ratings, ranking or recommendation"]}