What shipped
Dev/prod deploy scripts — separate deployment pipelines for staging and production environmentsCustom favicon and logo update — gave Fuscalendar its own visual identity"AI-estimated" rebrand — renamed the concept from "AI-generated tasks" to "AI-estimated tasks." The AI doesn't generate tasks, the user does. The AI just estimates how long they'll take. Words matterFrontend test suite — set up Vitest with @testing-library/svelte for component and integration testsBackend test suite — set up Bun's built-in test runner for API and service testsDaily estimation summary — added a time summary card to both the main page and the estimation review page, showing total hours with color-coded warnings (green under 6h, yellow under 8h, red over 8h)This was the maturity phase. No flashy features, just the kind of work that makes a project sustainable: tests, proper deployments, and a naming convention that actually reflects what the app does. The daily summary was the one user-facing addition, and it turned out to be one of the most useful, because seeing "Total: 10h" in red is a faster reality check than counting minutes in your head.