Context-free language (CFL) reachability is a widely used framework for formulating static program analyses. Operating over edge-labeled graphs, the standard algorithm performs context-free tabulation by iteratively deriving new edges that summarize paths whose labels conform to the production rules of a context-free grammar. However, as the term ``context-free'' suggests, these derivations are made without considering the surrounding context of inferred edges, often resulting in unproductive edges that do not contribute to the final reachability result.

In this paper, we present \emph{context-aware tabulation} (CAT), a novel approach that incorporates a restricted form of context sensitivity into CFL-reachability analysis to eliminate such unproductive edges. Comprehensive experiments on three widely studied clients show that CAT significantly accelerates reachability solving—achieving speedups of 1.75x, 1.49x, and 2.13x—and also reduces memory usage by 33.28%, 24.24%, and 50.15%, respectively, compared to a state-of-the-art solver.