Kinetic Canvas

Kinetic Canvas is an educational physics simulation that lets users explore fundamental physics concepts through interactive experimentation. Users can spawn particles, adjust gravity and friction parameters, and observe realistic collision detection and response in real-time. This project is part of the 777-1 Experiment, built with a "goldilocks zone" prompt that is intentionally underspecified to test how AI models handle ambiguity. After initial generation, 7 context engineer subagents (Amber, Kristy, Micaela, Lindsay, Eesha, Daniella, and Cassandra) will review the application in sequence, each focusing on their specialized domain. The findings will be documented in a case study that contributes to developing an algorithm for predicting prompt failures before they happen. As the most complex project in the experiment with no authentication requirements, Kinetic Canvas serves as the final test in the execution sequence, validating all patterns discovered in simpler projects.