Immature papers on arXiv
It seems to me that this question is less about the arxiv per se and more about how to navigate doing research in a very fast moving academic field.
I get the impression that arXiv is more than just putting something on a personal website.
It's certainly different. The main differences are:
(i) Many more people will see your paper.
(ii) Your paper will indeed be archived, essentially permanently. (Withdrawing a paper from the arxiv has the effect of uploading a new, empty version. Older versions are still there!) On your own website, you can take things down at least as quickly and easily as you can put them up.
(iii) Some (very obnoxious) journals may regard posting on the arxiv as "prior publication". (This is strictly unheard of in my field, mathematics. My guess is that CS is close enough to math so that it is at least very rare in yours.)
(iv) Minimum standards of completeness and professionalism are enforced on the arxiv. These are enumerated on the site itself, but the gist of it is that they are looking for manuscripts at the last step before conference/journal submission or later. They are not looking for early drafts.
Of these points, probably the last is most relevant to you. If it is standard in your subfield to include 25% experimental data [you say that is standard in "CS", but that is certainly not true across the entire field], then a paper uploaded to the arxiv without that would probably look to many in your field to be incomplete, which is against the spirit and perhaps the rules of the arxiv. So I wouldn't recommend it.
But the situation doesn't fundamentally change for papers that you or others post on your own website. The phrasing in your question suggests that you feel that you might not have to "be responsive" in the academic sense to papers that you find on people's webpages (only). That's not true. As an academic you have to be responsive to others' work wherever you find it.
In terms of the prospect of people uploading "the bare minimum to arXiv in order to get credit for the idea": is this an actual problem for you or just something you are wondering might be a problem? I have never encountered this problem in my work. That you are wondering whether it might be a problem makes me think you may be a quite new researcher and haven't fully grasped the way the academic community works. (Which is fine, and you have only to look forward to understanding it better. But you should talk to others, including advisors and mentors, to try to get a better idea.) Academia places a great privilege on completed work for exactly this reason. If you put out a manuscript which, say, modifies an algorithm and hints that it could be faster in some situations, the most likely reaction you'll get is "Go on..."
This question may finally have made me understand what people on this site are on about when they say things like "An idea is worthless". An idea is certainly not worthless, but a vague and unimplemented idea is of highly uncertain value, to the point where rushing to publish "only the vague idea" would be a very poor, um, idea.
By the way, you don't have to immediately drop something because someone else had "the same idea" and put out a paper before you. Much -- perhaps most -- important academic work overlaps with other work and even more of it refines and extends the ideas of others. How to respond to seeing "your idea" in another paper is a topic for a different answer.
Finally, let me say: if what you've done, are doing or want to do has real value, then it is unlikely to be received with thunderous applause this week and totally ignored next week. If you're living in fear that someone else will say what you want to say, maybe slow down and find more to say.
If there is a trend to upload just the bare minimum to arXiv in order to get credit for the idea.
No, there is not. No reputable author is stupid to publicly upload junk (on arxiv or anywhere else) and put his name on it. Perhaps, there are some people that do what you are suggesting (I cannot confirm though) but these mediocre-to-bad publications are nothing you should worry about.