<?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>CLI on Square Uncle</title>
    <link>https://squareuncle.com/en/tags/cli/</link>
    <description>Recent content in CLI on Square Uncle</description>
    <generator>Hugo -- 0.147.7</generator>
    <language>en</language>
    <lastBuildDate>Tue, 24 Mar 2026 12:45:00 +0800</lastBuildDate>
    <atom:link href="https://squareuncle.com/en/tags/cli/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Hand-rolled Lobster (3): Connecting WeChat to the CLI-Native Kernel</title>
      <link>https://squareuncle.com/en/posts/2026-03-24-visagent-wechat-clawbot-integration/</link>
      <pubDate>Tue, 24 Mar 2026 12:45:00 +0800</pubDate>
      <guid>https://squareuncle.com/en/posts/2026-03-24-visagent-wechat-clawbot-integration/</guid>
      <description>&lt;p&gt;&lt;img alt=&#34;VisAgent WeChat Integration Cover&#34; loading=&#34;lazy&#34; src=&#34;https://squareuncle.com/images/visagent_wechat_integration_cover.png&#34;&gt;&lt;/p&gt;
&lt;p&gt;Following our previous discussion on the &lt;a href=&#34;#ZgotmplZ&#34;&gt;RoleEngine Core&lt;/a&gt;, it&amp;rsquo;s time to bridge the gap between a raw CLI &amp;ldquo;brain&amp;rdquo; and a real-world communication platform: &lt;strong&gt;WeChat&lt;/strong&gt;.&lt;/p&gt;
&lt;p&gt;In VisAgent, we don&amp;rsquo;t believe in heavy, bloated frameworks. Instead, we use a &lt;strong&gt;CLI-Native Bridge&lt;/strong&gt; pattern. This post explores how we connected the &amp;ldquo;Clawbot&amp;rdquo; (our WeChat interface) to a hand-rolled Gemini CLI kernel.&lt;/p&gt;
&lt;h2 id=&#34;the-request-flow-from-chat-to-cli&#34;&gt;The Request Flow: From Chat to CLI&lt;/h2&gt;
&lt;p&gt;The architecture is a chain of specialized tools, each doing one thing well. Here&amp;rsquo;s how a message travels from your phone to the AI:&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
