Agent evals — cascivo
A published, re-runnable eval of how often Claude builds a working app with cascivo, with cascivo without its AI context, and with shadcn/ui: hidden Playwright tests, method and reproduction steps.
A published, re-runnable eval of how often Claude builds a working app with cascivo, with cascivo without its AI context, and with shadcn/ui: hidden Playwright tests, method and reproduction steps.