<?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/3fb5f74145e84f4dba342bd03dbd15c3&quot; frameborder=&quot;0&quot; width=&quot;2560&quot; height=&quot;1920&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1920</height><width>2560</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1920</thumbnail_height><thumbnail_width>2560</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/3fb5f74145e84f4dba342bd03dbd15c3-647c0515ed7ffd41.gif</thumbnail_url><duration>246.563</duration><title>Dyfram IR Tool - make.com Challenge - 04 Jun 2026</title><description>This Loom explains a workflow that converts a diaphragm governance core package into an implementation roadmap for AI builders. It describes a form that triggers a webhook to send data to a classification agent, which routes the package to one of seven agents that each use set algorithms, system prompts, and rack files to build a Google Docs file block by block until completion. After the roadmap is finalized, the system posts and emails informers of success and stores everything in a database, using Google Sheets for agent tracking and compliance. It also notes that invalid inputs route to an IR-2 failure sheet, and provides examples of the process and resulting emails.</description></oembed>