Preliminary diagnosis of closely resembling skin conditions can be highly subjective for dermatologists. In ambiguous cases, it often leads to performing invasive procedures like biopsies. Different skin conditions, however, have varying concentrations of fluorophores (like collagen, NADH) and chromophores (like melanin, hemoglobin) which can alter their fluorescence spectra. We demonstrate a handheld, portable, smartphone-based spectrometer that leverages these alterations in skin autofluorescence spectra for rapid screening of skin conditions. This methodology involves excitation of affected skin areas with ultraviolet (UV-A) 385 nm light, capturing the generated fluorescence spectra and sending the data wirelessly to a companion mobile application for data storage, analysis and visualization. By collecting the fluorescence spectral signals from healthy and unhealthy skin conditions, we establish that the signals collected using this portable device can be used to develop a classification method to help in differentially diagnosing these conditions. It shows promise as a useful skin screening tool for both dermatologists and primary health care workers. This device can enable quick, non-invasive and a more objective preliminary examination. We envision the device to be especially useful in primary healthcare centers of developing countries where availability of dermatologists is limited.