I’m building for the **Professional Agents** track. My project, **OpsGuardian**, is an autonomous incident-resolution agent that investigates operational failures, traces downstream impact, validates candidate fixes, and only asks for human approval when a repair crosses a defined risk boundary.
.png&w=256&q=75)