In any case, I don't think there is a law of nature or a biological force that makes women underestimate their skills. Plenty of women don't underestimate their skills.
It’s not like the job application/interview/etc dance is perfectly neutral towards every characteristic and background a person might have all the time as it is - words used, interview styles, and so on and so forth are never neutral.
I don’t know why studies suggest that boys are often better at correctly estimating (or maybe overestimating?) their skill level, but there’s lots of potential reasons aside from “biology” and “bad study”. Perhaps boys are just told “fake it til you make it” more often, and do so. Perhaps girls focus more on grades than boys in school for whatever reason, and the grading system is just set up to demoralise people in general. There’s a hundred thousand possible explanations - but they’re irrelevant to the point being made that if this is happening, then in order to encourage people who already believe these things about themselves to apply for a job, thinking about how to change your job posting to fit people with those beliefs is a good idea.
It is fine to try to attract more people to your job, if that is what you want. However, if the problem is "women underestimate their skills", perhaps it would be better to work on women having higher estimates of their skills, than on downscaling the demands of jobs.
Women are free to go to career counseling and learn to correctly estimate their skills. Like everybody else. Bookshops are full of books with career counseling, too. I bet you can even get free advice on the internet. If you want to, you can learn to succeed, even as a woman.
The whole discussion is moot. You are free to believe that women tend to underestimate their skills, and men tend to overestimate. It just happens to neatly align with feminist theory that men are worthless windbags, but if that is your world view, ok.
If you are looking for papers, you can find plenty that contradict your theory. For example today I saw another big study showing "Stereotype Threat" has no effect on girl's performance on Maths. So even if girls are told they are bad in Maths (which I doubt is happening on significant scale, but whatever), at the very least it doesn't affect their performance.