If you’ve had the experience of picking up software that has been poorly managed, a decision will confront you: fix it or ignore (work-around) it?
Better to fix it now.
The other option is of the assumption the poor code “isn’t that bad”and there are no consequences to stepping around the potholes and crafting yet more patches on patches. Until you are surprised to find that several months have gone by and you’re still pulling your hair out dealing with now really bad stuff.
If the fix is implemented immediately, the payback re-occurs every day from there on out. The old junk is no longer there, needing to be read, deciphered into some bizarre former context, then re-deciphered into a new context… Instead, you simply focus on the problem at hand in a productive manner.
(I’m not advocating wholesale re-writes; rather small incremental changes such as eliminating duplicated code, un-used strings, etc. These small bits of clutter, interfere with the bigger picture. And it keeps you in the practice of sharpening your skills.)
In my experience, you will be more productive by simply fixing the poor stuff immediately.