I’ve done LFS myself. After the core build, I moved on to BLFS to actually create something close to an everyday OS. I also tried to index the packages. But after a while, I got tired of all the work and moved on to other distro. I finally settled with Gentoo, with which I can customize the package compilation with way less effort.

LFS is a great tool for education & practice, but we couldn’t expect it to be productive as there are more effective & efficient tools out there that can do the things you do (and probably more) with way less time and effort.

One thing one would need to keep in mind is the great deal of effort to maintain the build on top of all the effort creating the build in the first place. If done by the book alone, it’d be like Sisyphus pushing a boulder uphill.

You can make automation tools for managing packages (or use automated LFS, part of the LFS project), but chances are the ones already out there can do better (e.g. apt, dnf, pacman, you name it). In fact, any tool you’d likely think of would’ve probably been created and tested.

With that said, you still can learn a great deal. Maybe it would give you all the knowledge and skill to do better, might even come up with something new.