Thought not a complete solution, I envision a github-like environment for the research and links to wikis and other repos for the lit.
People could openly submit "bugs" with methodology or conflicts between studies (as a replacement or supplement for peer review). You'd be encouraged as an author to publish research that was well followed/starred, but most importantly you'd be flagged if you had a bunch of open bugs.
(Bug resolution then would have to be approved by not just the researcher, but perhaps an Stack Overflow-like "close" voting system.)