<?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/0b27fcd1080048849de215764ddb6d0e&quot; frameborder=&quot;0&quot; width=&quot;1280&quot; height=&quot;960&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>960</height><width>1280</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>960</thumbnail_height><thumbnail_width>1280</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/0b27fcd1080048849de215764ddb6d0e-63343c4004993189.gif</thumbnail_url><duration>254.87</duration><title>Using AI Agents for Efficient Workflows</title><description>This Loom explains how to structure AI-assisted work into a clear four-step workflow and emphasizes making agent instructions concise and implementation-focused. The author describes drafting an “agent skill” that generates an implementation plan using terse sections such as summary, overview, approach, and numbered steps with links to specific code details and requirements. They run a separate “ImplementPlan” skill to carry out the work, noting that they are considering a rule to prevent editing the individual plan steps while implementation is in progress. They compare the readability of this newer plan to a previous, messier PR draft and conclude that the updated structure is easier to understand and covers key interfaces like the CLI.</description></oembed>