Team Idea
Coding agents don't get to say done, they prove it. When an agent finishes a change, it runs verify. The change goes to Vultr, where an independent agent on Vultr Serverless Inference writes its own test scenario, runs the app in a locked-down throwaway sandbox (web in a browser, Android on an emulator) and sends back screenshots, a recording, network and memory facts and a verdict. The agent fixes what failed before it reports back. Solo entry, Challenge 1: Blast Radius Zero.
Team Members
Leonidas Anagnou
leonidas_anagnou965