摘要： 论文阅读：Predict Before You Modify——让自进化 Agent 在修改自己之前先预测修改后果 论文标题：Predict Before You Modify: Learning Improvement Dynamics for Self-Evolving Agents 作者：Zi 阅读全文

摘要： 论文阅读：VERSE: Verified Self-Evolving Optimizer for Agent Harnesses 论文标题：VERSE: Verified Self-Evolving Optimizer for Agent Harnesses 作者：Zekai Wang, Yingq 阅读全文

摘要： 论文阅读：From Noisy Traces to Root Causes: Structural Trajectory Analysis and Causal Extraction for Agent Optimization 论文标题：From Noisy Traces to Root Caus 阅读全文

摘要： 论文阅读：Meta Context Engineering via Agentic Skill Evolution 论文标题：Meta Context Engineering via Agentic Skill Evolution 作者：Haoran Ye, Xuning He, Vincent A 阅读全文

摘要： 论文阅读：Inefficiencies of Meta Agents for Agent Design 论文标题：Inefficiencies of Meta Agents for Agent Design 作者：Batu El、Mert Yuksekgonul、James Zou 作者机构：Sta 阅读全文

摘要： 论文阅读：MetaSkill-Evolve——让 Agent 不仅进化 Skill，还进化“如何进化 Skill” 论文标题：MetaSkill-Evolve: Recursive Self-Improvement of LLM Agents via Two-Timescale Meta-Skill 阅读全文

摘要： 论文阅读：The Red Queen Gödel Machine——让 Agent 与 Evaluator 一起自进化 论文标题：The Red Queen Gödel Machine: Co-Evolving Agents and Their Evaluators 作者：Alex Iacob, A 阅读全文

摘要： 论文阅读：WebRL——通过自演化在线课程强化学习训练 LLM Web Agent 论文标题：WebRL: Training LLM Web Agents via Self-Evolving Online Curriculum Reinforcement Learning 作者：Zehan Qi, 阅读全文

摘要： 论文阅读：Agent Lightning: Train ANY AI Agents with Reinforcement Learning 论文标题：Agent Lightning: Train ANY AI Agents with Reinforcement Learning 作者：Xufang 阅读全文

摘要： 论文阅读：Understanding R1-Zero-Like Training: A Critical Perspective 论文标题：Understanding R1-Zero-Like Training: A Critical Perspective 作者：Zichen Liu, Chang 阅读全文