<?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/4d4c22353c5c4a1aacc7a95dab8cb698&quot; frameborder=&quot;0&quot; width=&quot;1794&quot; height=&quot;1345&quot; webkitallowfullscreen mozallowfullscreen allowfullscreen&gt;&lt;/iframe&gt;</html><height>1345</height><width>1794</width><provider_name>Loom</provider_name><provider_url>https://www.loom.com</provider_url><thumbnail_height>1345</thumbnail_height><thumbnail_width>1794</thumbnail_width><thumbnail_url>https://cdn.loom.com/sessions/thumbnails/4d4c22353c5c4a1aacc7a95dab8cb698-d58595085c8042a1.gif</thumbnail_url><duration>1224.734</duration><title>1. Install Contexer and Capture Decisions Automatically</title><description>This Loom explains how to install Contexer locally and use it to automatically capture and manage LLM agent decisions during a session. It describes running uv tool install Contexer followed by Contexer install, which sets up hooks for detected local agents so Contextel can start recording decisions as they occur. For the demo, the creator runs a cloud session with an initial no-context state, then has Claude implement a rate-limiting feature for a payment API in a repository and commit the changes. The Loom shows how captured decisions are initially pending, how they become authoritative after human approval, and how the decision details are stored for review and later sharing via a remote Contexer MCP.</description></oembed>