1 paper
Xue Jiang, Tianyu Zhang, Lingwei Wu +7
LLM agents have achieved strong performance on general software engineering tasks, yet struggle with domain-specific code generation. We identify the root cause as the agent's lack…