<?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>Middleware on AI Pulse</title>
    <link>https://cuigh.com/tags/middleware/</link>
    <description>Recent content in Middleware on AI Pulse</description>
    <generator>Hugo</generator>
    <language>en-US</language>
    <lastBuildDate>Wed, 03 Jun 2026 10:05:00 +0800</lastBuildDate>
    <atom:link href="https://cuigh.com/tags/middleware/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Why AI Agents Keep Failing: 90% of the Problem Is the Middle Layer</title>
      <link>https://cuigh.com/posts/agent-middleware-2026-github-middle-layer/</link>
      <pubDate>Wed, 03 Jun 2026 10:05:00 +0800</pubDate>
      <guid>https://cuigh.com/posts/agent-middleware-2026-github-middle-layer/</guid>
      <description>Looking at headroom, codegraph, and agentmemory topping GitHub Trending this week — and what they say about why AI Agents stall. The bottleneck isn&amp;#39;t the model. It&amp;#39;s the plumbing.</description>
    </item>
  </channel>
</rss>
