<?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/7b7a5dca86164d818e5517ed6e0114a5&quot; frameborder=&quot;0&quot; width=&quot;850&quot; height=&quot;637&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>637</height><width>850</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>637</thumbnail_height><thumbnail_width>850</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/7b7a5dca86164d818e5517ed6e0114a5-8ba856abc1d3897e.gif</thumbnail_url><duration>974.082</duration><title>Live Walkthrough of using Agent Scaffold  in a Project</title><description>This Loom explains how to use Agent Scaffold to maintain project context and governance across an agentic coding workflow. It describes an architecture with a governance framework plus a DuckDB-backed graph database that logs the project state, call points, modules, decisions, and findings semantically. The video walks through tools like Scaffold Orient to find what is in progress, draft, completed, and any blockers; Scaffold Impact and Scaffold Context to map file and API effects of changing a module; and Scaffold Prepare Review to run devil’s advocate adversarial reviews that challenge assumptions and feed gaps back into the plan. It also covers recording findings, surfacing open findings, tracing decision chains, beginning a plan with dependency checks, and comparing plans for overlap and recommendations on sequencing or parallelization.</description></oembed>