Nowadays, Serverless computing is emerging as one of the most used Cloud services. In particular, the Function as Service (FaaS) is bringing to Cloud consumers, developers, and devops many advantages in terms of service costs, speed of development, and ease of deployment. In fact, it stands to be a key technology for enabling the Cloud-Edge Continuum. Regardless of these features, it is still not possible to build FaaS native applications without a Cloud broker that coordinates the functions. Therefore, FaaS usage is limited to very simple and specific jobs. In this work, we brush up on the concept of Scientific Workflow using the FaaS paradigm, in order to realize full Native Serverless Workflows-based applications. We define a custom Workflow Manifest DSL used to describe function interactions, then we describe the implementation of an agent able to deploy architecture-independent functions and coordinate …