Let $b(n)$ denote the Euler quotient modulo $n$.
In OEIS we have A128465 Numbers k such that k divides the numerator of alternating Harmonic number H'((k+1)/2)
For $n>1$ we have $b(A128465(n))=2$.
Also all primes up to $10^8$ with $b(n)=2$ are in the sequence.
Q1 Is this relation between the sequence and Euler quotients true?
There are too few terms for experimental evidence.