video1.0<iframe src="https://www.loom.com/embed/a15bf3391c0945e7950ff213460d3ced" frameborder="0" width="1920" height="1440" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>14401920Loomhttps://www.loom.com14401920https://cdn.loom.com/sessions/thumbnails/a15bf3391c0945e7950ff213460d3ced-86218208802e691d.gif300.139This Loom explains the motivation and operation of a four agent system that monitors solar storms in near real time and issues live alerts for impacted infrastructure. The system reads live X-ray data from NASA GoEs Lite to detect solar flares, uses data between Earth and the Sun to detect expected wind arrival, and then identifies which infrastructure is likely to be impacted, supported by an LLM that provides reasoning. The author notes validation against events from about 21 years of history, referencing the 30 minute warning gap in a past case, and runs the pipeline for a storm about 96 hours beforehand to show warnings and affected assets. Alerts include satellite and high latitude risks, stating that satellites above 70 latitude fall into the “Rovo World Risk zone,” with examples showing high confidence up to 91% and issues around polar aviation routes and satellite activity.