The rapidly increasing complexity and scale of hybrid cloud environments requires improved service management capabilities in orchestration and automation. Current methods focus on provisioning infrastructure but lack functionality for consistently enabling and performing operational activities on managed services. We propose a data-driven approach to dynamically generate Orchestration Engine plugins from service descriptor metadata. Our approach extends Orchestration Engines by representing managed services as code within reusable blueprints in order to accelerate service deployments and ease management activities. In our work, we provide a data model and system architecture to allow service providers to easily author and publish resource definitions for a wide range of public and private services. These definitions may be combined into solution blueprints, forming a declarative and reusable representation of a managed workload. After provisioning a workload, administrators can view service instance data and invoke operational activities. For evaluation, we describe the authoring and orchestration of a hybrid cloud workload and discuss the strengths of our solution versus current methods.