---
id: 20260728-T0-02
title: "FlowEvo：让智能体通过工作流和技能协同进化，自动优化复杂任务"
title_en: "FlowEvo: Self-evolving agents through co-evolution of workflows and executable skills"
url: https://ai.daily.yangsir.net/daily/20260728-T0-02
issue_date: 2026-07-28
publish_date: 2026-07-27T04:00:00.000Z
category: research
source_name: "arXiv cs.AI"
source_url: https://arxiv.org/abs/2607.21596
---

# FlowEvo：让智能体通过工作流和技能协同进化，自动优化复杂任务

arXiv 上发布的新论文提出 FlowEvo 框架，让大模型智能体在执行复杂任务时，能自行发现并演化出有效的工作流程（推理、工具调用、代码执行）和可执行技能，而不只是依赖静态的推理方法。核心创新在于“工作流与技能的协同进化”，让智能体在任务求解过程中自动优化行为模式。

## English Version

**FlowEvo: Self-evolving agents through co-evolution of workflows and executable skills**

A new arXiv paper introduces FlowEvo, a framework enabling LLM agents to co-evolve workflows and executable skills during complex task execution. Rather than relying on static reasoning, agents autonomously discover and optimize inference-time procedures combining reasoning, tool use, and code execution.

---

**来源**：[arXiv cs.AI](https://arxiv.org/abs/2607.21596)

**详情页**：https://ai.daily.yangsir.net/daily/20260728-T0-02

---

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