Are you thinking of bed bugs? They don't dump their eggs, but have lower lifetime egg output because they are mitigating the attack. Seed beetles? Female gets injured, but the male's spines are linked to other traits that make her offspring do better. The net effect on her reproduction is actually positive despite the harm.

With ducks, the coercer quality hypothesis is an important hypothesis. But the evidence has caused the hypothesis terrain to shift to the indirect genetic benefit or reinforcement of existing preference hypothesis. Here's some situations that complicate the coercer quality hypothesis.

1. Coercers aren't reliably the "leftover" or inferior males as expect in the coercer quality hypothesis. It's mostly already-paired males doing the forcing, or, put another way, already screened and selected males which is arguably an indicator for genetic fitness.

2. If the tract exists to filter for good genes via female choice, you'd expect some positive relationship between choice-related investment and tract complexity. Instead, egg size (a proxy for female investment) is negatively correlated with both penis length and vaginal spiral count according to a 2020 study. This is the reverse of the prediction. Some in the research community argue female reproductive investment is doing more of the explanatory work. And not mate quality screening.

So the current leading hypothesis isn't "the tract detects and blocks bad genes," it's closer to "the tract protects a choice she already made through courtship, before force ever entered the picture." Coercion gets filtered because he isn't the one she picked. But this isn't settled yet and there are ongoing debate and more research to be done.

So your first paragraph is a bit irksome.

Probably because otherwise the females don't mate, or are killed during the mating or birthing process. Or it's possible that there are non-genetic factors to how they develop.

Because it's speculation to preserve the three explanations for all evolved traits that you know about.
(1) easier to survive -> killed during the mating or birthing process
(2) easier/more survivable reproduction -> otherwise the females don't mate (3) makes the mate more appealing so others want to mate with you.

Now hold the duck tract up against each one: Not (1). A corkscrew oviduct isn't a survival trait, it's purely reproductive plumbing. Not (3). It's not making anyone more appealing. It's not a peacock tail. Nobody's choosing a mate because of a more convoluted tract. Ipf anything it's invisible to mate choice entirely, it only matters after copulation. Not (2), and actually the opposite. The duck tract makes reproduction harder. It's a filter to block fertilization from certain partners.

And finally, we have enough evidence to show that this is a heritible trait. So what's going on?

This is called antagonistic coevolution. This is an evolved trait that makes it harder for all partners. The tract can relax for preferred partners making reproduction easier. But not as easy as a straight tract.

So we have convoluted tracts for the females... What do the males do? They create corkscrew penises. This goes back and forth, male adaptions force female adaptions. There is no directional selection of a fixed trait. There's coevolution based on frequency dependent selection. It's the Red Queen dynamics (that's the "it takes all the running you can do to keep in the same place" line from Alice in Wonderland). It's where male and female traits keep escalating against each other rather than converging on one fixed "best" trait. Rather you get a lot of opportunity for variability across the popultion and through generations. This is causal without being teleological.

Finally, there seems to be evidence that this can end in mutualism, but not guaranteed. I'm a little surprised you weren't more curious about this little factoid when I presented it and chose to close on your best understanding.

Anyway. I believe in unintelligent design.