1) The framework deliberately rejects industry-standard tooling like Composer for installation and uses a desktop app (actually an Electron app with a faux-retro theme) for installation, updates, and boilerplate code generation.
2) The framework itself is not written in idiomatic, modern PHP. It doesn't use namespaces, or strict_types, or indeed any types at all. It contains no inline documentation, but does (unfortunately) contain a substantial amount of inline HTML. There's no indication that any of it has been run through a static analyzer, or even a prettifier. I haven't looked at the code generated by the desktop app, but I have little doubt that it'd be similar.
3) There are some serious structural problems inherent to the framework. The most obvious is that their "Model" class isn't actually a model base class; it's a simple database abstraction layer which only supports trivial CRUD queries, and which conflates controllers with database tables.
4) The developers often seem focused on vanity metrics like gathering Github stars ("we need 1,200 GitHub stars to make Trongate a top ten PHP framework") and gaming benchmarks (requests/sec with an empty application is a poor metric).
None of this is conducive to a quality PHP framework. I don't want this to sound too harsh, but the authors really need to become more familiar with existing frameworks and development workflows before setting out to create their own.