A sample coverage rule accepts three timing errors above 10 minutes in 35 artificial AU Mic b transits in TESS Sector 95

Accurate transit times can constrain planets, but stellar activity can change the measured times. We tested a specified local transit fit on public AU Mic data from TESS Sector 95. We added 35 artificial b transits to the observed flux before normalization, baseline fitting, and flare masking. Each input used the fitted shape, a central depth of 2700 parts per million, and a known time offset of 15 minutes. A numerical rule checks positive depth, the time bound, and sample counts near transit entry, transit exit, and the baseline. It accepted 29 fits. Three accepted fits had time errors of +16.61, -39.19, and +24.43 minutes. Their formal errors were 0.35 to 0.47 minute. All 35 flat-flux controls at the same observation times passed with errors below 0.01 minute. A later audit of the three large errors used robust fits, cubic baselines, and free durations. All nine audit fits still had absolute errors above 10 minutes. Thus this coverage rule alone does not establish timing accuracy on these inputs. The windows can overlap. The 10-minute threshold and tail audit were selected after inspection. The trial counts are not a population failure rate. The test omits unknown regions on the stellar surface that a real planet covers. We also provide two real b timing estimates from 2025 with conditional scatter and method controls. We obtain no stable c time under the tested methods. The package includes original data, source records, code, results, figures, and offline reproduction steps.

AU Mic · TESS Sector 95 · transit timing · stellar activity · artificial transits · coverage rule · reproducible astronomy