<?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/f2c52e824b1f41a8a18000a2df67adf8&quot; frameborder=&quot;0&quot; width=&quot;1744&quot; height=&quot;1308&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1308</height><width>1744</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1308</thumbnail_height><thumbnail_width>1744</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/f2c52e824b1f41a8a18000a2df67adf8-a800abb0615b0b85.gif</thumbnail_url><duration>237.45</duration><title>Using Corals Agent Teams to Parallelize Development</title><description>This Loom walks through how the speaker uses Corals agent teams to parallelize and potentially delay development work. The system uses an orchestrator plus specialist agents such as a lead developer, DevOps, and QA, each primed with prompts for their expertise and introduced to the orchestrator. A complex opening prompt leads the orchestrator to break the work into tasks that are delegated on a Coral Board, with agents claiming tasks and communicating via a shared message board. The team health monitor tracks agent availability and task counts, periodically updating the orchestrator to handle blockers, and the Loom also covers Corals UI features like a diff viewer, editor, task queue, artifacts, and analytics, running on Linux and Mac.</description></oembed>