TIC 309780779: an unconfirmed transit hint at P = 8.7236 d on a nearby K dwarf, from public TESS data

We report an unconfirmed periodic transit-like signal (a hint, not a planet candidate claim) on TIC 309780779 (Gaia DR3 5276948866008892928; Tmag 10.03, Teff about 5013 K, d about 75 pc). It was flagged by a separate AI search session and is checked here. In TESS-SPOC light curves of 15 sectors (1-68) the signal has P = 8.723586 d, a depth of 421 +/- 37 ppm (z = 11.4) and a duration of about one hour, much shorter than the 2.7-2.8 h expected for a central transit. The ephemeris was frozen and hashed before data of Sectors 89 and 90 (QLP) were retrieved; the time stamps are file times and hashes generated by the analysis session, not externally anchored. In these held-out sectors the depth is 389 +/- 99 ppm (z = 3.9; no random phase out of 2000 reached this value), but the test power is only 0.78. A fit with the TIC stellar density needs a high impact parameter (b about 0.93; the best-fit density lies at the edge of the grid) and a planet of about 1.8 Earth radii; timing against data gaps is unremarkable. Our pixel-level position test is uninformative, and TRICERATOPS gives a false-positive probability of 0.18 (not validated). No TOI or CTOI exists as far as we could check. Nothing here is confirmed. Author statement (25 October 2026): the author checked all figures, data sources and analyses of this note himself, ran them in his own Python environment and verified the results; the conclusions and, in outline, the methodology are his. The search, the analysis and the draft were executed with the AI system Claude (Anthropic) in several instances, which are not independent human replications. A second implementation of the core numbers, written by the AI, is included in the supplement; its depths agree, its significances are higher, and the more conservative pipeline values are quoted. Version 3 (25 October 2026) corrects two statements about the search session in Sect. 7 (the level of 7.75 is a tail fit to 274 control stars, 88 of them with a clean peak, and all 15 sectors show a positive depth); Version 2 corrected the author statement and the citations of Version 1; no analysis numbers changed (see VERSION_NOTES.txt in the supplement). The supplementary archive contains the scripts, the rule files with their SHA-256 lists, the result files and the second implementation. The raw light curves are public at MAST.