<p>AGENTS.md is becoming the common language for AI coding tools, but keeping repo<br>
 rules, personal rules, and tool-specific files in sync is still messy. In this<br>
 episode, Kaushik and Iury break down the sync problem, compare their own setups,<br>
 and unpack what the latest AGENTS.md research actually says.</p>
<p><a href="https://fragmentedpodcast.com/episodes/306" rel="noopener noreferrer"><i>Full shownotes at fragmentedpodcast.com</i></a><i>.</i></p>
<h1>Show Notes</h1>
<h2>The sync problem</h2>
<ul>
 <li><a href="https://agents.md" rel="noopener noreferrer">AGENTS.md</a> - Official spec</li>
 <li><a href="https://developers.openai.com/codex/guides/agents-md/" rel="noopener noreferrer">Custom instructions with AGENTS.md</a> -<br>
  Open AI</li>
 <li><a href="https://kau.sh/blog/agents-md/" rel="noopener noreferrer">Keep your AGENTS.md in sync</a> - Kaushik's post</li>
 <li><a href="https://github.com/jpcaparas/rulesync" rel="noopener noreferrer">Rulesync</a> - What Iury uses</li>
 <li><a href="https://x.com/ryancarson/status/2024144258702143782" rel="noopener noreferrer">Tweet by Ryan Carson and Claude frustrations</a></li>
</ul>
<h2>Other links</h2>
<ul>
 <li><a href="https://arxiv.org/abs/2602.11988" rel="noopener noreferrer">Evaluating AGENTS.md: Are Repository-Level Context Files Helpful for Coding Agents?</a></li>
 <li><a href="https://openai.com/index/harness-engineering/" rel="noopener noreferrer">Harness engineering - Check the section about using AGENTS.md as a table of contents</a></li>
 <li><a href="https://opencode.ai/" rel="noopener noreferrer">OpenCode</a></li>
</ul>
<h1>Get in touch</h1>
<p>We'd love to hear from you. <a href="mailto:contact@fragmentedpodcast.com" rel="noopener noreferrer">Email</a> is the<br>
 best way to reach us or you can check our <a rel="noopener noreferrer">contact page</a> for other<br>
 ways.</p>
<p>We want to hear all the feedback: what's working, what's not, topics you'd like<br>
 to hear more on.</p>
<ul>
 <li><a href="https://fragmentedpodcast.com/contact" rel="noopener noreferrer">Contact us</a></li>
 <li><a href="https://buttondown.com/fragmentedcast" rel="noopener noreferrer">Newsletter</a></li>
 <li><a href="https://www.youtube.com/@fragmentedpodcast" rel="noopener noreferrer">Youtube</a></li>
 <li><a href="https://fragmentedpodcast.com" rel="noopener noreferrer">Website</a></li>
</ul>
<h2>Co-hosts:</h2>
<ul>
 <li><a href="https://kau.sh/" rel="noopener noreferrer">Kaushik Gopal</a></li>
 <li><a href="https://iurysouza.dev/" rel="noopener noreferrer">Iury Souza</a></li>
</ul>
<blockquote>
 <p>[!fyi] We transitioned from Android development to AI starting with<br><a rel="noopener noreferrer">Ep. #300</a>. Listen to that episode for the full story behind<br>
  our new direction.</p>
</blockquote>


<description>&lt;p&gt;AGENTS.md is becoming the common language for AI coding tools, but keeping repo&lt;br&gt;
 rules, personal rules, and tool-specific files in sync is still messy. In this&lt;br&gt;
 episode, Kaushik and Iury break down the sync problem, compare their own setups,&lt;br&gt;
 and unpack what the latest AGENTS.md research actually says.&lt;/p&gt;
&lt;p&gt;&lt;a href="https://fragmentedpodcast.com/episodes/306" rel="noopener noreferrer"&gt;&lt;i&gt;Full shownotes at fragmentedpodcast.com&lt;/i&gt;&lt;/a&gt;&lt;i&gt;.&lt;/i&gt;&lt;/p&gt;
&lt;h1&gt;Show Notes&lt;/h1&gt;
&lt;h2&gt;The sync problem&lt;/h2&gt;
&lt;ul&gt;
 &lt;li&gt;&lt;a href="https://agents.md" rel="noopener noreferrer"&gt;AGENTS.md&lt;/a&gt; - Official spec&lt;/li&gt;
 &lt;li&gt;&lt;a href="https://developers.openai.com/codex/guides/agents-md/" rel="noopener noreferrer"&gt;Custom instructions with AGENTS.md&lt;/a&gt; -&lt;br&gt;
  Open AI&lt;/li&gt;
 &lt;li&gt;&lt;a href="https://kau.sh/blog/agents-md/" rel="noopener noreferrer"&gt;Keep your AGENTS.md in sync&lt;/a&gt; - Kaushik's post&lt;/li&gt;
 &lt;li&gt;&lt;a href="https://github.com/jpcaparas/rulesync" rel="noopener noreferrer"&gt;Rulesync&lt;/a&gt; - What Iury uses&lt;/li&gt;
 &lt;li&gt;&lt;a href="https://x.com/ryancarson/status/2024144258702143782" rel="noopener noreferrer"&gt;Tweet by Ryan Carson and Claude frustrations&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;h2&gt;Other links&lt;/h2&gt;
&lt;ul&gt;
 &lt;li&gt;&lt;a href="https://arxiv.org/abs/2602.11988" rel="noopener noreferrer"&gt;Evaluating AGENTS.md: Are Repository-Level Context Files Helpful for Coding Agents?&lt;/a&gt;&lt;/li&gt;
 &lt;li&gt;&lt;a href="https://openai.com/index/harness-engineering/" rel="noopener noreferrer"&gt;Harness engineering - Check the section about using AGENTS.md as a table of contents&lt;/a&gt;&lt;/li&gt;
 &lt;li&gt;&lt;a href="https://opencode.ai/" rel="noopener noreferrer"&gt;OpenCode&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;h1&gt;Get in touch&lt;/h1&gt;
&lt;p&gt;We'd love to hear from you. &lt;a href="mailto:contact@fragmentedpodcast.com" rel="noopener noreferrer"&gt;Email&lt;/a&gt; is the&lt;br&gt;
 best way to reach us or you can check our &lt;a rel="noopener noreferrer"&gt;contact page&lt;/a&gt; for other&lt;br&gt;
 ways.&lt;/p&gt;
&lt;p&gt;We want to hear all the feedback: what's working, what's not, topics you'd like&lt;br&gt;
 to hear more on.&lt;/p&gt;
&lt;ul&gt;
 &lt;li&gt;&lt;a href="https://fragmentedpodcast.com/contact" rel="noopener noreferrer"&gt;Contact us&lt;/a&gt;&lt;/li&gt;
 &lt;li&gt;&lt;a href="https://buttondown.com/fragmentedcast" rel="noopener noreferrer"&gt;Newsletter&lt;/a&gt;&lt;/li&gt;
 &lt;li&gt;&lt;a href="https://www.youtube.com/@fragmentedpodcast" rel="noopener noreferrer"&gt;Youtube&lt;/a&gt;&lt;/li&gt;
 &lt;li&gt;&lt;a href="https://fragmentedpodcast.com" rel="noopener noreferrer"&gt;Website&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;h2&gt;Co-hosts:&lt;/h2&gt;
&lt;ul&gt;
 &lt;li&gt;&lt;a href="https://kau.sh/" rel="noopener noreferrer"&gt;Kaushik Gopal&lt;/a&gt;&lt;/li&gt;
 &lt;li&gt;&lt;a href="https://iurysouza.dev/" rel="noopener noreferrer"&gt;Iury Souza&lt;/a&gt;&lt;/li&gt;
&lt;/ul&gt;
&lt;blockquote&gt;
 &lt;p&gt;[!fyi] We transitioned from Android development to AI starting with&lt;br&gt;&lt;a rel="noopener noreferrer"&gt;Ep. #300&lt;/a&gt;. Listen to that episode for the full story behind&lt;br&gt;
  our new direction.&lt;/p&gt;
&lt;/blockquote&gt;
</description>

AGENTS.md is becoming the common language for AI coding tools, but keeping repo
rules, personal rules, and tool-specific files in sync is still messy. In this
episode, Kaushik and Iury break down the sync problem, compare their own setups,
and unpack what the latest AGENTS.md research actually says.

Full shownotes at https://fragmentedpodcast.com/episodes/306/

Fragmented - AI Developer Podcast

306 - Keeping your agent instructions in sync and effective

306 - Keeping your agent instructions in sync and effective

Description