this agent might perform better than OpenClaw and Hermes, just because of how it's built... most agent frameworks are one giant pile of code where everything is wired into everything else they're fast to start, but the moment it grows, one change breaks the whole thing and it takes a full day to fix it OpenSquilla did quite the opposite: > the core is tiny (around 100 lines) > its only job is to decide what happens and hand the work off > the model, the memory, the tools all live outside it…