<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>LLM Agent on AI内参</title>
    <link>https://www.neican.ai/tags/llm-agent/</link>
    <description>Recent content in LLM Agent on AI内参</description>
    <generator>Hugo</generator>
    <language>zh-cn</language>
    <lastBuildDate>Mon, 27 Oct 2025 10:40:04 +0800</lastBuildDate>
    <atom:link href="https://www.neican.ai/tags/llm-agent/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>LangChain 1.0核心迁移：从开源副业到AI Agent独角兽的架构深层进化与产业重塑</title>
      <link>https://www.neican.ai/insights/langchain-10ai-agent-20251027104004786-1/</link>
      <pubDate>Mon, 27 Oct 2025 10:40:04 +0800</pubDate>
      <guid>https://www.neican.ai/insights/langchain-10ai-agent-20251027104004786-1/</guid>
      <description>LangChain在获得1.25亿美元融资并估值达到12.5亿美元的同时，发布了其核心框架的1.0版本，这是一次基于LangGraph的彻底重写。此举通过统一Agent设计和引入强大的中间件机制，将LLM应用开发从“拼凑学”提升至工程化范式，深刻影响AI Agent的未来演进、人机协作模式及开源生态的商业化路径。</description>
    </item>
    <item>
      <title>A-MemGuard：解锁AI记忆黑盒，构建智能体时代的安全新基石</title>
      <link>https://www.neican.ai/insights/a-memguardai-20251016091004776-2/</link>
      <pubDate>Thu, 16 Oct 2025 09:10:04 +0800</pubDate>
      <guid>https://www.neican.ai/insights/a-memguardai-20251016091004776-2/</guid>
      <description>A-MemGuard创新性地解决了LLM智能体记忆投毒攻击这一关键安全问题，通过“共识验证”和“双重记忆”机制，有效识别并规避恶意推理路径，攻击成功率暴降95%以上。这项技术突破不仅提升了AI智能体的可信度和鲁棒性，更将加速其在金融、医疗等高风险领域的商业化落地，为AI安全伦理治理和未来AI社会构建了坚实基石。</description>
    </item>
  </channel>
</rss>
