This article presents a new paradigm for robust flutter modeling and analysis of high-order uncertain and linear aeroelastic systems. The fundamental idea is to couple the state-of-art in robust worst-case analysis (Linear Fractional Transformation modeling and µ analysis) with the state-of-practice in aeroelasticity (fluidstructure-interaction solvers). The issue with the latter is that, although capable of providing different levels of fidelity, they are less efficient in coping with the analysis of systems subject to uncertainties. In fact, while they have the advantage of capturing directly the physical uncertainty, the analyses can only be applied to a defined parameter combination, and due to their computational cost, it is usually only possible to consider a limited set of cases. To tackle this lack of robustness, in recent works the application of analytic worst-case methods has been proposed, but the intimately related problem of constructing accurate uncertain models has not been fully addressed. In this article, a co-modeling framework is presented that leverages the main features of both fluid-structure interaction solvers and robust control-based methods. The key idea is to combine these two typically distinct steps in a single one, enabling in this way to obtain an uncertainty description which is flexible and reconciles the physical sources of uncertainty with the uncertain parameters used in the LFT model. An exemplification of the developed framework on an unconventional aircraft configuration is provided. Results show its potential to provide valuable physical insights into the problem when analyzing complex systems.