(I may or may not have been involved with a company that did this to great success.)
The plugin today quite arguably scrapes Craigslist programmatically. I think Craigslist has a case there.
However, if Krrb instead injected JS at post-time to publish cross-site, I think that argument becomes much less convincing.
You can make a case on either side, but I think doing it at post time is important (for one thing, it guarantees that the publisher is actually the person krrbing it)
Edit to reply to below: Ah, ok. Your phrasing of "piggyback your plugin into popular installers to get this on millions of computers" sounded like you were planning to sneak it in without users' awareness. If it's an explicit opt in, then I would have no great problem with that from a user perspective. (Of course as a user I'd prefer to just install what I want, rather than being presented with options for a bunch of unrelated stuff, but advertising and affiliate programs keep things going, so I deal with it.)
Personally don't have a problem with the part about scraping something I might post on a classifieds site. Whenever I post something on a public web site I expect it to be scraped and to appear on search engines and such. It was specifically the idea of a service running on my computer, watching and potentially publishing aspects of my web use (and who knows what else), without my knowledge or permission that I took issue with.
I'm defining "piggyback downloads" as the extra page that comes up in someones installer. Isn't this a well-established practice? Maybe it's disappeared in the past few years?
I also love seeing this argument come within a community that's generally supportive of programmatically scraping that same user data and republishing it without user permission (although that was more what Padmapper did, not krrb).
Edit: your point re: increase chances of being sued vs. increase chances of success in court is well taken.