<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Ivan Wang</title>
    <description>The latest articles on DEV Community by Ivan Wang (@ivawzh).</description>
    <link>https://dev.to/ivawzh</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F46480%2Feb0d9215-82c2-4d68-b7d6-a5bc18758061.jpeg</url>
      <title>DEV Community: Ivan Wang</title>
      <link>https://dev.to/ivawzh</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/ivawzh"/>
    <language>en</language>
    <item>
      <title>Stop Fighting Your AGENTS.md File: A Better Way to Scale AI Agent Documentation</title>
      <dc:creator>Ivan Wang</dc:creator>
      <pubDate>Mon, 15 Sep 2025 14:39:31 +0000</pubDate>
      <link>https://dev.to/ivawzh/stop-fighting-your-agentsmd-file-a-better-way-to-scale-ai-agent-documentation-51n4</link>
      <guid>https://dev.to/ivawzh/stop-fighting-your-agentsmd-file-a-better-way-to-scale-ai-agent-documentation-51n4</guid>
      <description>&lt;p&gt;&lt;strong&gt;GitHub:&lt;/strong&gt; &lt;a href="https://github.com/ivawzh/agents-md" rel="noopener noreferrer"&gt;https://github.com/ivawzh/agents-md&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Your &lt;code&gt;AGENTS.md&lt;/code&gt; started innocent enough—maybe 50 lines of clean context. Six months later? It's a 5,000-line beast that nobody wants to touch.&lt;/p&gt;

&lt;h2&gt;
  
  
  The Problem Every AI Team Hits
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;project/
├── AGENTS.md     ← 5000+ lines, merge conflicts, stale info
├── docs/         ← team avoids updating the monster
└── api/          ← context scattered everywhere
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;&lt;strong&gt;Reality check:&lt;/strong&gt; Massive single files don't scale. Your team gets merge conflicts, information goes stale, and agents get overwhelmed with irrelevant context.&lt;/p&gt;

&lt;h2&gt;
  
  
  The Solution: Composable Agent Context
&lt;/h2&gt;

&lt;p&gt;What if you could organize agent context like code? With modules, clear boundaries, and automatic composition?&lt;/p&gt;

&lt;p&gt;&lt;strong&gt;agents-md&lt;/strong&gt; lets you split context into focused fragments, then auto-composes them into clean &lt;code&gt;AGENTS.md&lt;/code&gt; files wherever you need them.&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;project/
├── AGENTS.md                ← auto-generated, always fresh
├── docs/overview.agents.md  ← team docs agents can &lt;span class="nb"&gt;read&lt;/span&gt;
├── api/
│   ├── AGENTS.md            ← API-specific context
│   └── endpoints.agents.md  ← focused fragments
└── agents-md/epics/
    ├── epic-one.md          ← feature-specific context
    └── epic-two.md          ← easy to maintain
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;&lt;strong&gt;One command builds everything:&lt;/strong&gt;&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;npx agents-md compose
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;h2&gt;
  
  
  Key Features
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;
&lt;strong&gt;Smart targeting&lt;/strong&gt;: Fragments route to nearest &lt;code&gt;AGENTS.md&lt;/code&gt; automatically&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Zero config&lt;/strong&gt;: Works out of the box, customize only what you need&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Flexible organization&lt;/strong&gt;: Use any file structure that makes sense&lt;/li&gt;
&lt;li&gt;
&lt;strong&gt;Priority ordering&lt;/strong&gt;: Control what appears first with simple directives&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Get Started in 60 Seconds
&lt;/h2&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;&lt;span class="c"&gt;# Install and initialize&lt;/span&gt;
npm &lt;span class="nb"&gt;install&lt;/span&gt; &lt;span class="nt"&gt;-D&lt;/span&gt; agents-md &lt;span class="o"&gt;&amp;amp;&amp;amp;&lt;/span&gt; npx agents-md init

&lt;span class="c"&gt;# Create a fragment&lt;/span&gt;
&lt;span class="nb"&gt;echo&lt;/span&gt; &lt;span class="s2"&gt;"# API Guidelines..."&lt;/span&gt; &lt;span class="o"&gt;&amp;gt;&lt;/span&gt; docs/api.agents.md

&lt;span class="c"&gt;# Compose everything&lt;/span&gt;
npx agents-md compose
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;Done! Your &lt;code&gt;AGENTS.md&lt;/code&gt; now includes content from fragments and stays in sync automatically.&lt;/p&gt;

&lt;h2&gt;
  
  
  Why This Matters Now
&lt;/h2&gt;

&lt;p&gt;Every AI team hits this scaling problem. We're seeing explosive growth in agent projects, but tooling hasn't caught up. This is infrastructure every AI company will need.&lt;/p&gt;

&lt;h2&gt;
  
  
  Try It Today
&lt;/h2&gt;

&lt;p&gt;Fighting with unwieldy documentation? Give it a shot:&lt;br&gt;
&lt;/p&gt;

&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight shell"&gt;&lt;code&gt;npm i &lt;span class="nt"&gt;-D&lt;/span&gt; agents-md &lt;span class="o"&gt;&amp;amp;&amp;amp;&lt;/span&gt; npx agents-md init
&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;



&lt;p&gt;&lt;strong&gt;GitHub:&lt;/strong&gt; &lt;a href="https://github.com/ivawzh/agents-md" rel="noopener noreferrer"&gt;https://github.com/ivawzh/agents-md&lt;/a&gt;&lt;/p&gt;




&lt;p&gt;&lt;em&gt;What agent documentation pain points are you dealing with? Let me know in the comments or on GitHub!&lt;/em&gt;&lt;/p&gt;

</description>
      <category>ai</category>
      <category>agents</category>
      <category>documentation</category>
      <category>cli</category>
    </item>
  </channel>
</rss>
