<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Ai on No.Try not. Do… or do not. There is no try.</title><link>https://blog.vorona.me/tags/ai/</link><description>Recent content in Ai on No.Try not. Do… or do not. There is no try.</description><generator>Hugo</generator><language>en-US</language><copyright>Copyright © 2020–2022, Anatolii Vorona; all rights reserved.</copyright><lastBuildDate>Tue, 01 Sep 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://blog.vorona.me/tags/ai/index.xml" rel="self" type="application/rss+xml"/><item><title>scomp-mcp: Letting an AI Agent Use a Terminal on a Host It Cannot Reach</title><link>https://blog.vorona.me/posts/2026-09-01-scomp-mcp/</link><pubDate>Tue, 01 Sep 2026 00:00:00 +0000</pubDate><guid>https://blog.vorona.me/posts/2026-09-01-scomp-mcp/</guid><description>&lt;p&gt;I have a GPU host in a private network. It has no public IP and no SSH access&#10;from my laptop. I also have an AI coding agent on my laptop. The agent is good&#10;at builds and benchmarks, but it cannot reach that host.&lt;/p&gt;&#10;&lt;p&gt;The &lt;a href="https://blog.vorona.me/posts/2026-08-11-scomp-agent/"&gt;scomp agent&lt;/a&gt; already connects that&#10;host to the scomp relay with an outbound connection. So the missing part was&#10;small: a tool that lets a local model use the same terminal sessions, under the&#10;same rules. That tool is &lt;a href="https://github.com/smart-bit-me/scomp-mcp"&gt;scomp-mcp&lt;/a&gt;.&lt;/p&gt;</description></item></channel></rss>