Engineering Reliable Coding Agents: Evaluating and Operating the System Around the Model
构建可靠的编码智能体:评估与运行模型周边系统
专题命中 软件智能体 :coding agent(title,abstract);分类 cs.SE、cs.AI
AI总结 本研究针对AI编码智能体的可靠性问题,整合多源证据构建系统级评估与运行框架,区分模型与基础设施效应,提出可靠性记录目录及相关方法以提升智能体系统可靠性。
Comments Technical review and engineering monograph, 314 pages, 30 figures. Includes an evidence audit, a companion research artifact with 206 reliability records, and runnable protocols for evaluating and operating AI coding agents. August 2026. Source, companion, and reusable protocols: this https URL (https://github.com/sjarmak/engineering-reliable-coding-agents)