Twenty AU Mic events from 2025 give limited support for an early orbital phase direction under a sinusoidal rotation model

We test whether an orbital phase direction fit to AU Mic flare times from TESS sectors 1 and 27 agrees with sector 95. We reuse a fixed public 120-second flare search. A measured equivalent duration of at least 5 seconds selects 17, 19, and 20 events in 20.292, 19.106, and 18.150 days. The conditional event model includes the actual exposure and a separate sinusoidal rotation term for each sector. The early fit sets the orbital direction. In 2025, the signed efficient score along that direction is 1.754. Of 5,000 simulations under the fitted rotation-only model, 185 give a score at least this large. The one-sided simulation p-value is 0.0372, with Monte Carlo standard error 0.0027. The test log-rate amplitude is 0.584, with an approximate two-sided 95% profile interval of -0.053 to 1.306. All planned sensitivity intervals include zero. Combining two nearby peaks reduces the score to 1.522. Simulations have 48.2% power for a log-rate amplitude of 0.5, which gives a factor of 2.72 between the early maximum and minimum in the selected-event orbital term. The evidence gives limited support for agreement in this model. It does not establish a physical star-planet interaction. The study is retrospective, and the model assumes independent events and a simple rotation pattern. The package includes raw data, source records, code, figures, trial results, and reproduction instructions.

AU Mic · TESS · stellar flares · orbital phase · stellar rotation · star-planet interaction · conditional likelihood · reproducible research