1
Show HN: Veyra - Self evolving AI agent
自进化AI代理Veyra，从V0种子版本迭代升级，持续优化记忆与架构，打造自主演进系统
Article URL: https://github.com/iondodon/veyra Comments URL: https://news.ycombinator.com/item?id=49358474 Points: 1 # Comments: 0
2
3
AgentStream: How Well Do Self-Evolving LLM Agents Perform Under Streaming Tasks?
流式任务下自进化智能体表现如何？新基准AgentStream揭示性能边界与进化策略。
arXiv:2608.00155v1 Announce Type: new Abstract: Large language model (LLM) agents can self-evolve by continually improving from their own accumulated …
4
5
Kando AI完成数千万元种子轮融资，要做“决策领域的Cursor”｜涌现新项目
Kando AI获千万级种子轮，要做“决策界的Cursor”，专注提升决策采纳率与事后可靠性，AI从信息层迈向决策层。
文｜邓咏仪 编辑｜张雨忻 吴秉哲每天至少复盘一次。 作为北大计算机博士，他同时也是一个高频的投资者。每天盘后，他会回顾当天的判断——哪些预案执行了，哪些被盘面的新信息打乱了，哪些潜意识的决策事后被验证是对的。 这个习惯持续了很多年。但有一个问题是，大部分复盘都没被系统性地沉淀下来。 “你的认知是一种…
6
不换模型，效果提升104%！上海AI Lab让Harness也能自进化了
不换模型，只改Harness就能让AI性能飙升104%？上海AI Lab自我进化新方案，效率惊人！
Harness本身也可以被搜索、验证和迭代
7
From Atomic Actions to Standard Operating Procedures: Iterative Tool Optimization for Self-Evolving LLM Agents
从原子动作到标准流程，让LLM智能体自主迭代优化工具，突破固定策略天花板。
arXiv:2607.07321v1 Announce Type: new Abstract: Tool utilization enables Large Language Model (LLM) agents to interact with the real world and resolve…
8
翁荔新博客提出「自进化先从Harness开始」，DeepSeek崔添翼转发附议
翁荔提出自进化从Harness而非模型权重开始，元智能体优化Agent工作流设计，崔添翼转发支持
崔添翼：这个方向很容易出成果
9
SkillCoach: Self-Evolving Rubrics for Evaluating and Enhancing Agentic Skill-Use
让AI学会自我评估技能运用，动态进化评分标准，精准提升智能体能力。
arXiv:2607.01874v1 Announce Type: new Abstract: Skills are becoming a reusable operational layer for LLM agents, encoding SOPs, domain rules, tool wor…
10
AgRefactor: Self-Evolving Agentic Workflow for HLS Compatibility and Performance
AI智能体自动重构HLS代码，兼顾兼容性与性能，自进化工作流让硬件设计效率起飞。
arXiv:2606.30949v1 Announce Type: new Abstract: High-Level Synthesis (HLS) provides a fast path from concepts to silicon, but converting real-world so…
11
Self-Evolving World Models for LLM Agent Planning
让智能体在规划中自我进化世界模型，论文提出全新训练范式，值得关注。
arXiv:2606.30639v1 Announce Type: cross Abstract: World models offer a principled way to equip long-horizon LLM agents with foresight: predictions of …
12
SEVA: Self-Evolving Verification Agent with Process Reward for Fact Attribution
打破验证黑箱，SEVA用过程奖励让大模型幻觉可追溯、可自纠。
arXiv:2606.29713v1 Announce Type: cross Abstract: Hallucination is the reliability bottleneck for LLM-based agents, and fact attribution verifiers are…
13
Agentic Hardware Design as Repository-Level Code Evolution
HORIZON将硬件设计转化为仓库级代码进化，AI智能体框架实现硬件自进化，颠覆传统设计流程。
arXiv:2606.28279v1 Announce Type: cross Abstract: We present HORIZON, a self-evolving agent framework that treats hardware design as repository-level …
14
Safety in Self-Evolving LLM Agent Systems: Threats, Amplification, and Case Studies
自进化LLM智能体暗藏系统性安全风险？这篇论文系统梳理威胁与放大机制，并给出实证案例验证。
arXiv:2606.23075v1 Announce Type: cross Abstract: Self-evolving LLM agent systems, which autonomously update their model parameters, memory, tools, an…
15
Self-Evolving Multi-Agent Systems via Textual Backpropagation
提出基于文本反向传播的多智能体自我进化框架，让智能体在协作中自动优化策略，无需人工干预。
arXiv:2506.09046v3 Announce Type: replace Abstract: Leveraging multiple Large Language Models (LLMs) has proven effective for addressing complex, high…
16
Memory Beyond Recall: A Dual-Process Cognitive Memory System for Self-Evolving LLM Agents
突破传统记忆局限：双过程认知记忆系统让LLM智能体实现自进化学习。
arXiv:2606.09483v1 Announce Type: new Abstract: Long-term memory for an LLM agent is more than retrieving the right passage at the right time. Current…
17
Socratic-SWE: Self-Evolving Coding Agents via Trace-Derived Agent Skills
提出自进化编码代理Socratic-SWE，通过轨迹推导技能突破SWE任务训练瓶颈，为LLM代理能力提升开辟新路径。
arXiv:2606.07412v1 Announce Type: cross Abstract: LLM-driven software engineering agents have become a central testbed for real-world language-model c…
18
The Self-Evolving Agent: How to Build Closed-Loop AI Systems That Write and Optimize Their Own Code
让AI自动编写并优化自身代码的闭环系统实践指南，从提示工程到生产自适应性，破除“一次性完美”迷信。
We have all been there. You spend hours meticulously crafting the perfect system prompt or tool description for your AI agent. It performs beautifully…
19
SkillRevise: Improving LLM-Authored Agent Skills via Trace-Conditioned Skill Revision
LLM Agent技能自进化新方法：利用轨迹条件修正，解决冷启动下初始不完美技能改进难题。
arXiv:2606.01139v1 Announce Type: new Abstract: Agent skills are procedural artifacts that enable LLM agents to execute workflows, verify constraints,…
20
Harness Updating Is Not Harness Benefit: Disentangling Evolution Capabilities in Self-Evolving LLM Agents
LLM智能体进化研究新视角：解耦外部框架更新与性能收益，揭示自进化能力的真正来源
arXiv:2605.30621v1 Announce Type: new Abstract: LLM agents are increasingly deployed as systems built around editable external harnesses, including pr…