<?xml version="1.0" encoding="UTF-8"?><oembed><type>video</type><version>1.0</version><html>&lt;iframe src=&quot;https://www.loom.com/embed/dbb4a56b6a934fc1b19664d8b7240d61&quot; frameborder=&quot;0&quot; width=&quot;1920&quot; height=&quot;1440&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1440</height><width>1920</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1440</thumbnail_height><thumbnail_width>1920</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/dbb4a56b6a934fc1b19664d8b7240d61-1b71a7b48123594d.gif</thumbnail_url><duration>555.217</duration><title>How to Walk Test Your Brain Setup</title><description>This Loom explains how to walk test a Brain or ClaudeMD routing and context setup before going live. Rymac shows that the router (ClaudeMD) holds about 35 lines, maps job types to business folders per line, and contains no content, while a separate context MD describes what is true right now and includes decisions and rules, including not needing to re-ask “who I am.” He demonstrates doing a walk test in a cloud environment by reading only the ClaudeMD file in the target folder, requiring the model to name the correct folder and next file and to quote the line that justifies routing. A pass is defined as naming the folder, file, and check using the start file and one context MD without opening a tree, while a fail occurs when it asks where something is, opens the wrong folder, or reads a tree first.</description></oembed>