I imagine it is fairly common for students (at the PhD level) to produce work of lower quality than their advisors. Indeed, even excellent students must start somewhere, and if they were able to produce work of the quality of professors, then they need not be a student. Moreover, not all students will be exceptional. At the same time, it is often the case that papers of a student would be co-authored by the advisor. How are such papers are evaluated by the broader community. Does taking on students who publish papers of lower quality (relative to the advisor) impact the advisor's reputation in any meaningful way (e.g. by "diluting" the quality of their work)?
While the practice of adding advisor's name to student publications varies by field, as does the interpretation of it, the advisor in such fields has no incentive to have low quality work published with their name as "author" (whatever that might mean).
Thus, an advisor will, for their own purposes, push the student to do more so that the quality is high.
You are likely correct that most students, at that moment, are less skilled in research and publication than their advisors, but that is only at the moment. If it were generally true over time then the quality of research would decline as time goes on and nothing significant can be done any more. Person A (good) produces student B (less good) who produces student C (sort of bad), etc until all is garbage. Only Euclid's work in math would be worth reading.
Also, the advisor has the opportunity to feed ideas to their students, which can be a justification for advisor authorship in the first place. Thus, the quality can be as good as what the advisor could produce - at that moment.
My dissertation was a fine piece of work. I was congratulated by members of my defense committee on its quality and completeness. My advisor (an eminent, senior mathematican) could have done it himself, certainly, but he expressed pride in my work. This being math, he wasn't a co-author though the idea for the original exploration was his. He had an idea that "such and so might be true", which is IMO the highest form of math insight and I hadn't attained that yet. But doing the work led me to that level. The overall trend is upward, not downward.
So, if you ask how such work is judged in general, then the answer is that it is judged by the research community like any other. On its merits.
But the judgement by the university for matters of promotion and salary advancement is probably a bit different. In that case, more is better, I suspect.
No. For a supervisor to coauthor papers with a student, which are lower quality than the supervisor's solo work, does not harm the supervisor's reputation.
The researchers judged to be most successful publish both quality and quantity. Publishing somewhat lower quality research with a student contributes to the quantity.
Publishing in disreputable journals or presses will harm the supervisor's reputation.
From an ethical point of view, routine and unremarkable research provides great value to society even when it does not get judged a "success" by the research community. Much technological achievement comes from boring, systematic experimentation.