MOSAIC picks a GraphRAG traversal policy per query and beats the best fixed policy by 9.96 points.

arXiv 2609.11065 starts from a structural mismatch: direct facts want compact local neighborhoods, comparisons want balanced coverage of several targets, mediated questions need deeper paths through weak connectors, and most systems traverse identically for all three. Mosaic is training-free. An LLM analyzer converts each query's evidence requirements into a bounded policy over seed selection, traversal, stopping and evidence selection, while graph, indexes, scoring and generator stay fixed. On GraphRAG-Bench it reaches 76.97 query-weighted Answer Correctness on Medical and 64.33 on Novel, 5.13 and 4.43 points over prior best. The 9.96-point gap against the strongest fixed policy is the number that matters if you already run GraphRAG.