Research data support in libraries is moving into the mainstream and libraries are stepping up their activities and involvement with research disciplines in various ways. The core business of libraries is typically to provide support on how to access research data and related infrastructures and to advise on good data management practices (e.g. data management planning). Based on this, libraries may consider stepping up their actual involvement in working with research data. However, this comes with new challenges in terms of available data (science) skills and capacities. A workshop at LIBER's 2018 Annual Conference was organised by the Research Data Management Working Group, in order to address the following questions: What kind of data science skills do we need in libraries? (in the context of research support) What approaches have been taken to step up capacity and skills? How successful are training approaches like DS4L, Carpentries (Data, Software), etc.? Are they sustained and expanded after the training? Additionally, the workshop will provide a report about recent activities of the working group, e.g. a library of data management plans and the collection of case studies.