基于智能体原生可复用代码原语的大规模仓库工程
原标题：Large-scale Repository Engineering via Agent-Native Reusable Code Primitives
AI 摘要
论文提出 Code Primitives（智能体原生的可复用可执行组件），每个组件自带接口契约、依赖闭包、验证测试和来源信息，并可通过常驻 LLM 作为自然语言接口适配目标仓库；这些组件被组织成可检索库 CodeFace。作者进一步提出 LEGO 流水线，通过检索激活、组件协作、诊断验证三机制协调组件完成仓库级构建，并发布 LEGO-REPO 基准（覆盖七个软件领域、五个难度等级）。实验显示 LEGO 提升了所有被测模型，GPT-5.6-terra 的交付分数显著上升，且适配后的组件优于直接检索上下文或原样引入的代码。
正文节选
Large-scale Repository Engineering via Agent-Native Reusable Code Primitives Abstract Large language models (LLMs) equipped with development environments have moved code generation toward repository-scale construction, yet building complete repositories remains difficult because interacting modules, interfaces, configurations, tests, and dependencies must work together. We introduce Code Primitives, agent-native reusable executable components that carry their own interface contracts, dependency