In this work, we propose partial RNA design, a novel RNA design paradigm formulated as a constraint satisfaction problem. Partial RNA design describes the problem of designing RNA sequences from arbitrary RNA sequence and structure motifs. Our formulation enables RNA design of variable-lengths candidates, while still allowing precise control over sequence and structure constraints at individual positions. Based on this formulation, we improve an existing inverse RNA folding algorithm with a masked training objective to derive libLEARNA, a new algorithm capable of efficiently solving different tasks in the realm of RNA design. A comprehensive analysis across various problems, including a realisitic riboswitch design task, reveals the outstanding performance of libLEARNA and its robustness.