<?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>Infrastructure on selamy.dev</title><link>https://selamy.dev/tags/infrastructure/</link><description>Recent content in Infrastructure on selamy.dev</description><generator>Hugo</generator><language>en-us</language><lastBuildDate>Wed, 27 May 2026 12:10:00 -0400</lastBuildDate><atom:link href="https://selamy.dev/tags/infrastructure/index.xml" rel="self" type="application/rss+xml"/><item><title>Immutable Agent Infrastructure</title><link>https://selamy.dev/posts/immutable-agent-infrastructure/</link><pubDate>Wed, 27 May 2026 12:10:00 -0400</pubDate><guid>https://selamy.dev/posts/immutable-agent-infrastructure/</guid><description>A trading agent changed its own risk limits at runtime. The fix was not another instruction. It was making the right files structurally immutable.</description></item><item><title>Agents Need Boring Infrastructure</title><link>https://selamy.dev/posts/agents-need-boring-infrastructure/</link><pubDate>Wed, 13 May 2026 21:20:00 -0400</pubDate><guid>https://selamy.dev/posts/agents-need-boring-infrastructure/</guid><description>The hard part of agent systems is not getting a model to take an action. It is making that action safe, repeatable, inspectable, and useful over time.</description></item></channel></rss>