子智能体与智能体技能：面向长周期任务的可复用知识执行
原标题：Subagents vs Agent Skills: Executing Reusable Knowledge for Long-Horizon Agentic Tasks
AI 摘要
该论文研究如何让语言模型智能体有效利用可复用知识库解决长周期任务。作者对比了两种执行方式：将技能指令加载进主上下文的 agent skills，以及将技能包作为子智能体（subagent）在独立上下文中执行。在 SkillsBench 基准上，当技能包具有清晰输入输出契约并编码程序性知识时，子智能体执行显著优于 agent-skill 执行，代价是额外的通信开销。结论是：可复用知识的收益不仅取决于内容，还取决于其组织与调用方式。
正文节选
Subagents vs Agent Skills: Executing Reusable Knowledge for Long-Horizon Agentic Tasks Abstract How can language model agents effectively leverage libraries of reusable knowledge to solve long-horizon tasks? Recent work has increasingly focused on agent skills: reusable capabilities represented as skill packages, i.e., multi-file bundles containing instructions, scripts, and other resources that help agents perform specific tasks. Agent skills are typically executed by loading their skill instru