<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>Context Gateway on 赛博工具站</title>
    <link>https://haodaohang.top/tags/context-gateway/</link>
    <description>Recent content in Context Gateway on 赛博工具站</description>
    <image>
      <title>赛博工具站</title>
      <url>https://haodaohang.top/images/cover.png</url>
      <link>https://haodaohang.top/images/cover.png</link>
    </image>
    <generator>Hugo -- 0.152.2</generator>
    <language>zh-cn</language>
    <lastBuildDate>Sat, 14 Mar 2026 14:00:00 +0800</lastBuildDate>
    <atom:link href="https://haodaohang.top/tags/context-gateway/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>100万token的诅咒：Agent上下文膨胀与压缩解法</title>
      <link>https://haodaohang.top/posts/2026-03-14-agent-context-compression/</link>
      <pubDate>Sat, 14 Mar 2026 14:00:00 +0800</pubDate>
      <guid>https://haodaohang.top/posts/2026-03-14-agent-context-compression/</guid>
      <description>GPT-5.4支持100万token上下文，但测试显示准确率从32k的97.2%暴跌到1M的36.6%。本文探讨Agent上下文管理的核心矛盾与智能压缩解法</description>
    </item>
  </channel>
</rss>
