@ryanc I was actually thinking whether some (not so) fancy crypto could be used to pass some instead of a bool that the attacker can't forge, then realized reverse proxy configs are not exactly designed to implement such transformations in the first place :)
Nonetheless, this is an illustrative example that unless we point to some robust solution ppl *will* come up with complex but insecure solutions (see also Schneier's Law).
@screaminggoat@zeljkazorz@GossiTheDog "However, due to inadequate server configurations, attacks become possible if *the serialized data is not verified* (CWE-642)" - this sounds more like disabled MAC than leaked key to me
@GossiTheDog 1) 3000 is not a big number on the Internet (quality matters though) 2) This is an overestimation because not all keys are useful (as the captured text also implies)
I haven't touched ASP.NET for a while, but I'd risk to say that app configuration also affects exploitability as i) not all apps rely on signed ViewState (IIRC) ii) deserialization gadgets are not universal.
These are of course solvable problems, but still need to be taken into account for risk assessment.
@GossiTheDog That is technically true, but scanners already look for exposed web.configs, so any affected, but not already exploited Internet-facing sites would be simultaneously extremely negligent and lucky.
- They come without all the invasive crap of smart phones - They boost creativity - They teach user interfaces and controls outside "push shiny moving button" - They teach basic software concepts like files (yes, knowing about files is a skill) and how to move them around
"One enduring complication with all this is that scraping happens all the time for reasons that people *don’t* find inherently objectionable, and in fact support—the Wayback Machine, all kinds of public health and extremism research, etc. The mistake was assuming that goodwill transfers.
A key problem in the Disc Horse (and policy to a lesser extent) is reminding people that scraping as a technological process is Important, Actually, for all the things You Think Are Good, and any proposed solutions to curtail GAI training uses need to be VERY narrowly tailored to not impact those.
All the proposed solutions so far have had some critical flaw that makes them unworkable.
Manual consent? Ok, how do we implement that at scale? robots.txt style flags are fine, but they’re also not legally binding—and that’s good! If they were, Wayback wouldn’t be able to index!
So exclusion protocols can be ignored, For Good Reason. “What if we give an exclusion protocol the force of law for this specific use?” Closer, but there’s active debate in the courts about whether this is all a fair use, and if the answer is “yes,” then it doesn’t matter
…then best case scenario the tags are rendered null (because you can’t legally override fair use), and worst case you’ve just recreated a DMCA 1201 style lockout trick, and we have spent the last 25 years seeing just how incredibly those fuck up everything around them."
@GossiTheDog@reverseics@cR0w But could *non-admin* users access the DB of *other* users? SQLite or not, this should not be possible (in general...). If it was possible back then (as it was suggested by you and articles based on your comments), then now would be the best time for all to see what the problem was to check if the same or similar problem is present in the implementation that is to be released.