The iTask SVG Image library uses a combinator approach to specify the layout of graphical elements. These graphics can be made interactive, by associating callback functions with images. The type system is used to make sure that the callback functions operate on view-model values of the correct type. A long-standing open problem with the library is how to deal with interactive images that work on different view-model values. This is a highly desirable feature because it allows code reuse and modularization. A solution should not ‘pollute’ the set of image combinators with view-model plumbing combinators. It should also not alter the design philosophy that considers images as being layers of images on top of a host image. We propose a solution to this problem by recursively applying editor task customization to images.