I was inspired to ask this problem after trying to find all $(x,y,u,v)$ for which $xy+1,xu+1,xv+1,yu+1,yv+1,uv+1$ are all sqaure.
After some basic calculation I was easily able to find $x=n, y=n+2, u=4n+4$. However, I had some difficulty finding $v$.
It took some time, but I was able to find $v=4(n+1)(2n+1)(2n+3)$.
Are there more easier ways of finding $(x,y,u,v)$?
And is the following generlization possible: There exists infinitely many sets $X_{ n }=\{ x_{ i }|1\le i\le n \quad i\in \mathbb{N}\}$ and ${ X }_{ n }\subset \mathbb{N}$ for any finite natural number $n$ for which ${ x }_{ i }x_{ j }+1$ are all perfect squares?
This problem is well-known in the literature, where such sets are called Diophantine $4$-tuples. Dujella has made an extensive study of them: see his web page at https://web.math.pmf.unizg.hr/~duje/dtuples.html for plenty of references.
In particular, Dujella proved that there are no Diophantine $6$-tuples and that any $5$-tuples, if they exist, are bounded by an absolute constant so there are at most finitely many (none have been found). So the conjecture in your last paragraph is false for every $n \ge 6$.
Your question about extending a triple of integers to a $4$-tuple has also been considered (see https://web.math.pmf.unizg.hr/~duje/quint.html for the reference). We may always take
$$v = x + y + z + 2xyz + 2\sqrt{(xy+1)(yz+1)(xz+1)}.$$
It's conjectured that the above value of $v$ is unique (which would immediately imply that $4$ is the largest possible Diophantine tuple).