We study a nonparametric search problem to detect an anomalous stream in a finite set of S data streams. The anomalous stream is an i.i.d sequence drawn from the distribution q, while the other data streams are i.i.d sequences drawn from a distribution p. The distributions p and q are assumed to be arbitrary and unknown, but distinct. We propose a universal, distribution-free sequential test that is exponentially consistent and stops in finite time almost surely. We also derive the limiting growth rate of the expected stopping time as the probability of error decreases to zero. Simulations show that the performance of the proposed test is at par or better than that of the fixed sample size test for this setting. Our test is the first such universal sequential test for the hitherto open general nonparametric case of unknown, arbitrary distributions.