---
id: 20260911-T0-13
title: "子代理 vs Agent技能包：哪种方式让LLM更擅长长周期任务？"
title_en: "Subagents vs Agent Skills: Which Helps LLMs Execute Long-Horizon Tasks Better?"
url: https://ai.daily.yangsir.net/daily/20260911-T0-13
issue_date: 2026-09-11
publish_date: 2026-09-10T04:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2609.09233
---

# 子代理 vs Agent技能包：哪种方式让LLM更擅长长周期任务？

一篇arXiv论文对比了子代理（Subagents）和Agent技能包（Agent Skills）两种方案在执行长周期任务时的效果。技能包将可复用能力打包为多文件束，而子代理则通过分工协作完成任务。研究探讨了语言模型代理如何有效利用可复用知识库来应对复杂的长周期任务，为Agent系统设计提供了参考。

## English Version

**Subagents vs Agent Skills: Which Helps LLMs Execute Long-Horizon Tasks Better?**

An arXiv paper compares subagents and agent skills for executing long-horizon tasks. Agent skills package reusable capabilities as multi-file bundles, while subagents divide work through coordinated delegation. The research examines how language model agents can effectively leverage libraries of reusable knowledge to solve complex, long-horizon tasks, offering design insights for agent systems.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2609.09233)

**详情页**：https://ai.daily.yangsir.net/daily/20260911-T0-13

---

*智语观潮 · Daily — https://ai.daily.yangsir.net/llms.txt*