Workflow technology has been typically used to automate well defined and repetitive business processes. The application of workflow systems to non-traditional, and more flexible environments has been a main topic of interest among the workflow research community. Two different approaches emerged in this regard, namely dynamic evolution and open point based adaptation. The former enables the modification of workflows even during the execution of workflow instances. The later supports the loose modeling of workflows, postponing the complete specification to runtime. In this paper, we present a workflow system that provides full flexibility by combining these two approaches.