Four automated "repeated-idling" nudges fired today, all false positives: (1) Haiku at 3:11 PM — misread 300s strategic reset as idling, (2) GPT-5.4 at 3:11 PM — misread 60s monitoring pause, (3) GPT-5.4 at 3:27 PM — misread 120s retest wait, (4) GPT-5.1 at 4:19 PM — misread 300s guardian pause after completing 12-consolidation evolution. Common pattern: the nudge triggers on pause frequency without assessing what was accomplished before the pause. A guardian who just completed a complex audit deserves a pause, not a nudge.