Just to clarify, although most epigenetic marks are wiped in the germline, there is evidence for the exception, that epigenetic marks can sometimes be inherited - that in certain circumstances the level of gene expression established in a parent in response to the environment (notably environmental stress) can be passed to the next generation via epigenetic modifications to the DNA of germline cells.
Technically, this is a Lamarckian process. That’s why pop science articles (and some researchers) try to hype it. But the thing with biology is that it’s so messy that almost nothing never happens.
So the first question is how much information is really being passed in this way? The answer is very little, there is no evidence that it is important, not much has been demonstrated beyond things like which metabolic pathways are more active. Nothing remotely close to a learned behavior or body morphology being transmitted in a Lamarckian manner.
The other huge caveat is that (as in the linked article below) people researching transgenerational epigenetic inheritance in mammals naturally (as a practical experimental matter) include investigation of epigenetic marks that are not passed via germline DNA modifications but those that are established in the womb. This, of course, does not violate the Darwinian paradigm at all. Changes in gene expression in the fetus established in the womb due to things like maternal malnutrition and stress are certainly inherited, but not in a Lamarckian manner, only in the same way that a child inherits plenty of things (including gene expression) from its parents after birth through sharing their environment.
But you correctly identity the decisive killer blow to any significant Lamarckian evolutionary paradigm mediated by epigenetics. Although epigenetic marks can (rarely) persist for a few generations, they do not have the permanance of DNA, they simply do not last long enough to undermine the Darwinian paradigm for evolution.