<?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>Agents on Code &amp; AI</title>
    <link>https://jupaygon.dev/en/tags/agents/</link>
    <description>Recent content in Agents on Code &amp; AI</description>
    <image>
      <title>Code &amp; AI</title>
      <url>https://jupaygon.dev/images/og-default.jpg</url>
      <link>https://jupaygon.dev/images/og-default.jpg</link>
    </image>
    <generator>Hugo</generator>
    <language>en</language>
    <copyright>2026 Juanjo Payá</copyright>
    <lastBuildDate>Fri, 04 Sep 2026 00:00:00 +0000</lastBuildDate>
    <atom:link href="https://jupaygon.dev/en/tags/agents/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Agents as users of the infrastructure</title>
      <link>https://jupaygon.dev/en/posts/2026-09-04/agent-as-a-user/</link>
      <pubDate>Fri, 04 Sep 2026 00:00:00 +0000</pubDate>
      <guid>https://jupaygon.dev/en/posts/2026-09-04/agent-as-a-user/</guid>
      <description>What your agents actually do on your machines, if you let them. Watch out for the sudoers that grants root, the SSH allowlists that let them hop anywhere, and other hazards of trusting them too much.</description>
    </item>
    <item>
      <title>Loop engineering: designing the system that prompts the agent</title>
      <link>https://jupaygon.dev/en/posts/2026-06-17/loop-engineering/</link>
      <pubDate>Wed, 17 Jun 2026 00:00:00 +0000</pubDate>
      <guid>https://jupaygon.dev/en/posts/2026-06-17/loop-engineering/</guid>
      <description>Addy Osmani and Boris Cherny gave it a name: you stop prompting the agent and start designing the loop that prompts it. Here are their six pieces, which I&amp;#39;ve spent months trying out in my own setup, plus some critical small print.</description>
    </item>
  </channel>
</rss>
