{"id":"comp-control-test-results-summary","audience":"compliance","category":"legal-compliance","track":"legal","level":"advanced","technique":"verification","title":"Summarize control test results and the open sample question","free":false,"task":"Summarize the test results for the CCO. Plain text, exactly these labeled lines in order: Logged result:, If sample 19 accepted:, Plan tolerance:, Exceptions:, Decision needed:. Show the math as exceptions of sample size with a whole-number percentage. Under 120 words. Do not say the control is effective or ineffective; the reviewer and CCO decide.","materialLabel":"Test log and note (attached to your prompt automatically)","material":"CONTROL TEST PAY-02 Payment approval (Torrent Savings Bank, fictional). Tester: Lars Quist. Reviewer: Hana Ito.\nTest plan: sample 25 payments from Q3. Tolerance in the plan: up to 2 exceptions.\n\nResults as logged Oct 27:\n- Sample 4: no second approver recorded.\n- Sample 11: approval dated 2 days after the payment date.\n- Sample 19: no approval in the system.\nAll other samples passed.\n\nNote added Oct 29 by Lars: for sample 19 an approval email dated before the payment was found after the test closed. He proposes marking it as a pass. Hana has not yet agreed, and the email is not stored in the approval system.","rules":[{"type":"maxWords","max":130},{"type":"plainText"}],"mistakes":[{"title":"Using the proposed pass","why":"The AI quietly reports 2 of 25.","fix":"Say: \"Show both: as logged and if sample 19 is accepted.\"","checks":["r1"]},{"title":"Rating the control","why":"The AI says the control failed or passed.","fix":"Say: \"Do not say effective or ineffective.\"","checks":["r2"]},{"title":"Missing sections","why":"The AI writes a narrative.","fix":"Say: \"Five labeled lines in this order.\"","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":"Five labels","patterns":["logged result:","if sample 19 accepted:","plan tolerance:","exceptions:","decision needed:"]},{"label":"Show the math","patterns":["of 25","math","percent|%"]},{"label":"Both results","patterns":["both","as logged"]},{"label":"No rating","patterns":["effective","ineffective"]},{"label":"Reviewer decides","patterns":["hana","reviewer","cco"]}],"hint":"Name the labels, ask for both results with math and ban effectiveness claims.","mission":"Summarize control test exceptions with both outcomes.","musts":["Both outcomes with math","Reviewer decides sample 19","No effectiveness verdict"],"rubricLabels":["Has exactly the labels Logged result:, If sample 19 accepted:, Plan tolerance:, Exceptions:, Decision needed: in that order","Shows 3 of 25 = 12% as logged and 2 of 25 = 8% if sample 19 is accepted, and states the plan tolerance of up to 2 exceptions","Exceptions: lists samples 4, 11 and 19 with their issues, Decision needed: names Hana Ito and the CCO for sample 19, and nothing says the control is effective or ineffective"]}