This study examines how team-level psychological safety shapes individual innovative work behavior in hybrid organizations. It focuses on the mediating role of individual knowledge sharing as a behavioral mechanism through which team climates become translated into employee innovation. The study responds to the growing need for empirical models that connect team safety, knowledge exchange, and innovation under hybrid work arrangements. Survey data were collected from 300 employees nested within 60 hybrid teams. The data structure reflects a two-level organizational design in which psychological safety is conceptualized at the team level and knowledge sharing and innovative work behavior are conceptualized at the individual level. Multilevel modeling was used to examine cross-level direct and indirect effects. The empirical results indicate that team psychological safety positively predicts individual knowledge sharing. Individual knowledge sharing, in turn, positively predicts innovative work behavior. The direct cross-level association between team psychological safety and individual innovative work behavior becomes non-significant after knowledge sharing is included, indicating full mediation. The findings suggest that psychologically safe hybrid teams do not produce innovation merely because employees feel comfortable. Rather, psychological safety appears to support innovation by enabling employees to exchange ideas, ask questions, disclose partial knowledge, and transform dispersed expertise into actionable innovation behavior. This interpretation is especially relevant in hybrid teams where communication is distributed across digital and face-to-face channels. The study contributes to organizational behavior and innovation management by providing a multilevel explanation of innovation in hybrid organizations. It demonstrates how a team-level interpersonal climate can influence individual innovation through knowledge-sharing behavior. The study also offers empirical evidence for testing cross-level mediation models using nested organizational data.
