This paper describes a framework for identifying, classifying, and coding student proofs, modified from existing proof-grading rubrics. The framework includes 20 common errors, as well as categories for interpreting the severity of the error. The coding scheme is intended for use in a classroom context, for providing effective student feedback. In developing the framework, we applied previously published rubrics to college-level student proofs, and then added, merged, or split categories based on the frequency of the most common proof errors. We tested the new framework on student proofs from a variety of undergraduate courses, and found that it reasonably captured student errors. The framework, its design and limitations, and code descriptors are provided.