Hand gesture pertains to certain positions or movements of the hand that communicate a distinct meaning. Typically, in the realm of research, these two contexts refer to static and dynamic gestures. In the scope of this paper, we are focusing exclusively on static gestures, which are defined as fixed or held positions of the hand, as opposed to dynamic gestures that involve continuous movement. Hence, all references to gestures in this paper imply static gestures only. Hand gesture recognition is often relevant in various situations. Possible applications include sign language identification and dance gesture recognition, among others. Most existing studies on dance gesture recognition have focused on image-based features, which struggle to generalize across different dance forms due to inconsistencies in style and context. Moreover, most of the works employed either handcrafted features or deep learning-based approaches but not both. In this work, we are addressing these challenges specifically on the recognition of hand gestures of Sattriya dance, a traditional dance form originating from Assam, India. We are utilizing skeleton-based information and combining both handcrafted and deep learning features to enhance recognition accuracy. This integrated approach leverages the strengths of both feature types and shows promising results in accurately classifying Sattriya dance hand gestures, advancing the field beyond traditional image-based methodologies.