Among granular matter, one type of particle has special properties. Upon being assembled in disordered configurations, these particles interlock, hook, almost braid, and – surprisingly, considering their relatively low packing fractions – show exceptional shear strength.Such is the case of non-convex particles. They have been used in the shapes of tetrapods, ‘L’, ‘Z’, stars, and many others, to protect coasts or build self-standing structures requiring no binders or external supports. Although these structures are often designed without a comprehensive mechanical characterization, they have already demonstrated great potential as highly resistant construction materials. Nevertheless, it is natural to attempt to find the most appropriate non-convex shapes for any given application. Can a particle shape be tuned to obtain a desired mechanical behavior? Although this question cannot be answered yet, current technological, simulation, and experimental developments strongly suggest that it can be resolved in the next decade. A clear understanding of the relationships between particle shapes, mechanical response, and packing properties will be key to providing insights into the behavior of these materials. Such work should stand on 1) robust and general shape descriptors that encode the complexity of non-convex shapes (i.e., the number of arms, the symmetries, and asymmetries of the bodies, the presence of holes, etc.), 2) the analysis of the response of assemblies under different loading conditions, and 3) the disposition and reliability of non-convex shapes to ensure durability. The manufacturing process and an efficient use of resources are additional elements that could further help to optimize particle shape. In the quest of designing bespoke non-convex particles, this paper consolidates the challenges that remain unresolved. It also outlines some routes to explore based on the latest developments in technology and research.