Detection catalog and diagnostic plots for "Shrinking the Haystack: One-Class Machine Learning Detection of Magnetosheath Current Sheets in MMS Burst Data"

This dataset contains the 270 retained current-sheet detections from a one-class Deep SVDD anomaly detection pipeline applied to 15 magnetosheath turbulence intervals (1.58 h of MMS1 burst-mode data, 2015-10 to 2016-12), as described in Ghuge et al. (2026), "Shrinking the Haystack" (arXiv:2608.00252). For each detection this includes an 11-panel diagnostic plot, a per-detection time-series CSV, and a metadata JSON file sufficient to reproduce the detection window, along with the manual visual-screening labels (candidate reconnection / sheet-like / not convincing) summarized in Table 2 of the paper.