<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Androsmeda Constellation</title><link>https://blog.androsmeda.com/</link><description>Recent content on Androsmeda Constellation</description><generator>Hugo</generator><language>en</language><lastBuildDate>Sun, 08 Nov 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://blog.androsmeda.com/index.xml" rel="self" type="application/rss+xml"/><item><title>Building the operator: an AI agent that holds the SSH keys, and the six layers that hold it</title><link>https://blog.androsmeda.com/posts/building-the-operator/</link><pubDate>Sun, 08 Nov 2026 00:00:00 +0000</pubDate><guid>https://blog.androsmeda.com/posts/building-the-operator/</guid><description>&lt;p>I run two AI agents at home. The &lt;strong>assistant&lt;/strong> reads my documents, mail and the web, keeps notes and drafts things. The &lt;strong>operator&lt;/strong> is the only one allowed to touch my servers and network. This post is how the operator is built, layer by layer, with the actual mechanisms. Names and addresses are placeholders; the shapes are real.&lt;/p></description></item><item><title>About</title><link>https://blog.androsmeda.com/about/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://blog.androsmeda.com/about/</guid><description>&lt;p>I build and run a home lab: a storage server, a few small machines, and a growing set of AI agents that help run it. This blog is where I write down how it goes, for other home-lab people, for anyone curious about letting AI agents near real systems, for friends who keep asking what all those blinking boxes do, and for myself, so I remember why I did things.&lt;/p></description></item></channel></rss>