Non-Civil Servant or non-Aparatur Sipil Negara (non-ASN) contract personnel are a distinct public-sector Human Resource Management (HRM) group because they perform routine government work while remaining outside the formal ASN career structure and face greater contract uncertainty than Civil Servants or Pegawai Negeri Sipil (PNS) or Government Employees with Work Agreements or Pegawai Pemerintah dengan Perjanjian Kerja (PPPK) employees. This study examines the association between organizational culture, leadership style, work motivation, and self-rated employee performance among non-ASN contract personnel at the Regional Secretariat or Sekretariat Daerah (Setda) of Batu Bara Regency, North Sumatra, during the post-2024 employment status transition during the employment-status transition following Law No. 20 of 2023 and KepmenPANRB No. 16 of 2025. A cross-sectional survey was conducted with 56 usable responses from a population of 132 non-ASN personnel. The instrument used seven-item Likert scales for each construct, with item-total correlations exceeding the critical r = 0.263, and Cronbach's alpha values ranging from 0.889 to 0.948. Multiple linear regression showed that organizational culture (beta = 0.403; p < 0.001), leadership style (beta = 0.508; p < 0.001), and work motivation (beta = 0.469; p < 0.001) were each positively associated with self-rated employee performance, and the three predictors were jointly significant (F(3, 52) = 32.83; p < 0.001; adjusted R2 = 0.635). Leadership style emerged as the strongest correlate, suggesting that supervisor behavior is especially salient in a hierarchical bureaucracy where contract personnel depend heavily on immediate supervisors for task direction, recognition, and perceived job security. Because the data were cross-sectional, single-source, and self-rated, the findings were interpreted strictly as context-specific associations rather than causal effects. This study contributes initial evidence on an understudied non-ASN workforce and recommends multi-source, longitudinal, and multi-regency replications.