A faint dip in 2025, found again in 2018: an unconfirmed 2.5 Earth-radius planet candidate around the active K dwarf TIC 257484419

In 2025 the TESS satellite recorded a small, regular dip in the light of a faint, active orange dwarf 80 parsecs away. The dip repeats every 4.3533 days. Because the timing is so regular, it can be run backwards, so I did that: the ephemeris predicted a transit in 2018, in data from the first year of the mission. Five independent reductions of those old images show a dip at the predicted times. If it is a planet, it is about 2.5 times the size of the Earth, circles its star in 4.35 days at 0.045 AU, and gets roughly 57 times the starlight that we do. It is an unconfirmed candidate, not a discovery. The star is spotted and may be young, its Gaia astrometry is slightly noisy, and I have no high-resolution images or spectra. A close companion could still be the source. A statistical check gives a false-positive probability of about 0.09, which is "likely planet", not "validated". The paper lists everything I tested, what could still be wrong, and what would settle it (mainly high-resolution imaging). Everything comes from public TESS and Gaia data. The files: the preprint, a one-line ephemeris table, six figures, and an artist's impression of the planet (an illustration, not data).