WELDON is aimed at enhancing the autonomy of bipedal robots, such as humanoids, for future application to manufacturing, service robotics, and healthcare domain. Recent robotic challenges have shown that humanoids exhibit serious robustness problems within unstructured scenarios. On the other hand, robot manipulation capabilities are daily increasing, filling the actual gap between human and robotic skills. Latest funded projects in robotics have demonstrated that robots can perform complex manipulation tasks like the nonprehensile ones, where the object is manipulated without being caged within the hand’s fingertips or the palm. A new trend in robotics is to find out possible connections between different research fields. By using existing techniques in a specific robotic sector, it is then possible to export the discovered tools into the connected scenario to solve some issues. To this purpose, literature has already highlighted that balancing, slow walking gaits, and grasping tasks share several similarities. For instance, two slow walking legs can be seen as two fingers grasping a much larger object, i.e., the earth. However, walking shares also some hybrid contact/non-contact conditions with a multi-fingered hand juggling some objects on it. Since the robotic manipulation theory is indeed advanced during the last decades, the fundamental idea behind WELDON is to establish a connection between multi-contact dynamic, energy efficient walking and nonprehensile manipulation. Find out such a connection will allow to WELDON to go beyond the current scientific state of the art in legged robotics, providing both theoretical and technological solutions where cutting-edge knowledge and innovation can merge in creating e versatile framework for walking tasks.