Molecular gut content analyses (MGCA) are employed to study diets without microscopy or stable isotope analysis, providing far higher taxonomic resolution than is possible with traditional methods. Most MGCA are done with metabarcoding, but whole-metagenome sequencing (WMGS) has the potential to provide even greater taxonomic specificity than metabarcoding, making it a powerful method for food-web studies. Despite its advantages, the use of WMGS for MGCA has pitfalls, such as the inclusion of false positives, contamination of reference databases, failures of accepted tools, improper application, computational limitations, and conflicting biological signals. Our study tests the potential of shotgun sequencing-based MGCA to inform stream food-web dynamics by using Pycnopsyche, a common genus of shredding caddisfly found in forest streams in the eastern U.S.A., as a model organism. We collected Pycnopsyche larvae from a developing cohort (second- to fifth-instar) in a single headwater stream at the Coweeta Hydrologic Laboratory, North Carolina, U.S.A. DNA from the caddisfly guts was extracted and prepared libraries sent for WMGS. Preliminary results show promising potential, but these techniques should be adopted with caution. Using the DNA extracted from Pycnopsyche’s gut contents, we highlight potential insights and problems associated with the methodology.