Exploring new meshing technologies and the ecosystem of tools around them has been, and will mostly remains, a field of research aimed at improving any numerical simulation whether in terms of accuracy, performance, or robustness of the solutions. Over the years, research on numerical methods for polytopal meshes have gained a lot of traction for different scientific communities. But despite this growing body of publication depicting that polytopal methods are advantageous in numerous ways, the meshing aspect of polytopal simulation remains largely open to improvements. This work introduces a general framework for simulation-driven adaptive polytopal meshing based on Centroidal Voronoï Tessellations (CVT). Two modifications are introduced to the classical CVT formulation to make it suitable for simulation-driven adaptation. First, h-adaptivity is achieved through a controlled seeding strategy : starting from a coarse CVT mesh, the distribution of Voronoi generators is improved according to a density function derived from simulation error indicators, so that regions exhibiting larger solution variations receive a higher generator density. Second, r-adaptivity is achieved by modifying the standard Lloyd iterations to compute weighted generator displacements instead of uniform centroid updates within the classical CVT. The weights are again constructed from solution-dependent quantities using a tailored FEM-like discretization within each polytopal centroid computation. Together, these two extensions preserve the favorable geometric properties of CVT meshes while enabling effective, fully simulation-driven mesh hr-adaptation. The proposed methodology is tested through numerical benchmarks and designed to be generic and scalable, leveraging current open-source tools for Vorono¨ı-based computations within our algorithms. Alongside this work, we are planning in the future to release an open-source implementation of those algorithms in the PRISM framework, providing the community with a practical and extensible tool for hr-adaptive polytopal meshing.