Congenital scoliosis and congenital vertebral malformations arise from heterogeneous disruptions of vertebral segmentation and formation, leading to a highly variable clinical morphology and associated anomalies. Over the past decade (2015-2025), human genetics and developmental biology have clarified core etiologic pathways, particularly somitogenesis/segmentation clock-wavefront signaling and TBX6-associated congenital scoliosis, while supporting multi-factorial contributions in many apparently sporadic cases. This review integrates morphology with spatiotemporal embryologic windows, linking axial patterning (e.g. RA-HOX context), epigenetic regulation, segmentation clock signaling (Notch/WNT/FGF), and downstream tissue-scale effectors to vertebral segmentation defects and three-dimensional curve phenotypes. We propose an evidence-grading scheme (levels A-D) to differentiate well-supported mechanisms and clinically actionable genetic findings from emerging hypotheses and preclinical signals. The synthesis highlights practical implications for molecular diagnosis and variant interpretation, phenotype-informed test selection, and counseling, while outlining priority directions for resolving non-coding/structural variation, gene-environment interactions, and mechanistic validation.