Robot skins have emerged recently as products of research from various institutes worldwide. Each robot skin is designed with different applications in mind. As a result, they differ in many aspects from transduction technology and structure to communication protocols and timing requirements. These differences create a barrier for researchers interested in developing tactile processing algorithms for robots using the sense of touch; supporting multiple robot skin technologies is non-trivial and committing to a single technology is not as useful, especially as the field is still in its infancy. The Skinware middleware has been created to mitigate these issues by providing abstractions and real-time acquisition mechanisms. This article describes the second revision of Skinware, discussing the differences with respect to the first version.