<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Engineering on Notes from the Rabbit Hole</title><link>https://magnus919.com/categories/engineering/</link><description>Recent content in Engineering on Notes from the Rabbit Hole</description><generator>Hugo</generator><language>en</language><copyright>© [Magnus Hedemark](https://github.com/magnus919)</copyright><lastBuildDate>Mon, 18 May 2026 14:58:50 -0400</lastBuildDate><atom:link href="https://magnus919.com/categories/engineering/index.xml" rel="self" type="application/rss+xml"/><item><title>Architecture by Leverage: The Five-Sentence Design Philosophy That Changed How I Build Agents</title><link>https://magnus919.com/2026/05/architecture-by-leverage-the-five-sentence-design-philosophy-that-changed-how-i-build-agents/</link><pubDate>Mon, 18 May 2026 14:58:50 -0400</pubDate><guid>https://magnus919.com/2026/05/architecture-by-leverage-the-five-sentence-design-philosophy-that-changed-how-i-build-agents/</guid><description>&lt;p>I saw something in the Hermes Agent Discord that stopped me mid-scroll.&lt;/p>
&lt;p>Not a feature announcement. Not a breaking change notice. Not a philosophical debate about agent architectures. Just someone — user &lt;code>sovthpaw&lt;/code> — describing how they actually use their agent in five sentences:&lt;/p>
&lt;blockquote>
&lt;p>If you upload your Obsidian knowledge base to your Google Workspace, you can version control, share, and backup your knowledge base.&lt;/p>
&lt;p>My agent GitHub integration is largely centered around an &lt;code>/llm-wiki&lt;/code> (or &lt;code>/obsidian&lt;/code> works) in my agent&amp;rsquo;s own dedicated repo.&lt;/p></description></item><item><title>The Hermes Kanban: A Complete Guide to Multi-Agent Task Orchestration</title><link>https://magnus919.com/2026/05/the-hermes-kanban-a-complete-guide-to-multi-agent-task-orchestration/</link><pubDate>Mon, 18 May 2026 00:00:00 +0000</pubDate><guid>https://magnus919.com/2026/05/the-hermes-kanban-a-complete-guide-to-multi-agent-task-orchestration/</guid><description>&lt;h1 id="the-hermes-kanban-a-complete-guide-to-multi-agent-task-orchestration">The Hermes Kanban: A Complete Guide to Multi-Agent Task Orchestration&lt;/h1>

&lt;div class="info-box">
 &lt;div class="info-icon">
 &lt;svg xmlns="http://www.w3.org/2000/svg" width="24" height="24" viewBox="0 0 24 24" fill="none" stroke="currentColor" stroke-width="2" stroke-linecap="round" stroke-linejoin="round">
 &lt;circle cx="12" cy="12" r="10">&lt;/circle>
 &lt;line x1="12" y1="16" x2="12" y2="12">&lt;/line>
 &lt;line x1="12" y1="8" x2="12.01" y2="8">&lt;/line>
 &lt;/svg>
 &lt;/div>
 &lt;div class="info-content">
 &lt;strong>Ahead of the release:&lt;/strong> This guide covers the kanban system as it will be once PRs &lt;a href="https://github.com/NousResearch/hermes-agent/pull/27572">#27572&lt;/a> (orchestrator-driven auto-decomposition) and &lt;a href="https://github.com/NousResearch/hermes-agent/pull/27813">#27813&lt;/a> (artifact delivery) are merged. Both have been reviewed and approved by the Hermes maintainers. If you&amp;rsquo;re reading this before they land, your kanban board has the specify tool but not the decompose tool, and &lt;code>kanban_complete&lt;/code> doesn&amp;rsquo;t yet accept the &lt;code>artifacts&lt;/code> parameter — everything else applies today.
 &lt;/div>
&lt;/div>

&lt;p>If you&amp;rsquo;ve used Hermes Agent for more than a weekend, you&amp;rsquo;ve probably discovered &lt;code>delegate_task&lt;/code>. It&amp;rsquo;s the obvious tool when you need a subagent to handle a subtask: fork work to a child, wait for it to return, collect the result. It looks like a function call. It works like a function call. And for one-shot reasoning subtasks, it&amp;rsquo;s the right tool.&lt;/p></description></item></channel></rss>