I think it would be pretty interesting to work at a place where they the expectation of being defrauded is high enough that they actually have a line item on the budget for it.
It all comes down to risk appetite. The assumption is that there will be x amount of fraud losses. Mitigation by y% will cost z. They do the maths and come up with an answer of "No, we won't bother to fix this application, because we only expect £100,000 losses this year and it will cost £100,000 to fix
Interesting statistic I heard in training last week: An un-patched Windows system can be compromised within 20 minutes of being connected to the Internet. From a fresh install, it takes at least 30 minutes to download all necessary Windows updates.
Your trusted perimeters end up being more tightly restricted.
Without going into potentially redactable information from public record laws, a lot of universities end up having a "soft" perimeter at the campus network ingress.
Then having designating network segments, such as data centers, HR departments, as hard perimeters.
@RoryAlsop Students are the worst attackers: they have time, they have huge available computing resources, and they are ready to do attacks for the sake of it, even without any rational incentive.
I suggest checking all their windows for "the algorithm". That's where you need to look for students causing problems (haven't you seen the social network???!!).
I would like to know if it is save for the host system of a virtual machine (VM - VirtualBox OSE in my case) to execute malware.
Can a virus break out and read or write data from the host system? Can it establish an internet connection if I disable it in my VM?
Is a VM a safe environment to tr...
@Ninefingers Not completely, in my opinion. The main theme (can code break out of a VM ?) is a dupe of past questions; but the bit about fork bombs is new.
Right, weekend time for me. @AviD take a look at my github repository, let me know what you think. For a few hours work in my spare time I have a driver that might even be able to communicate /w userland; gimme a bit of time at the weekend and we might even be able to capture Irps
then after that, I need registry and network monitoring
@thisjosh Yes, absolutely none of the French nuclear plants would resist a coordinated attack by 3000 allosaurus driving battle tanks (with laser turrets).
I've just discovered that my application contains a security vulnerability for other accounts on the local system... I'll describe the problem and then if anyone would like to offer a suggestion, that would be great :)
So I have written an application that runs on a user's local account.
An amusing thing is that the environmentalists and the other parties fight over the question about whether closing nuclear plants will create or destroy jobs
...however, the problem is that the server is still available to other users on the same machine and they can access information that should be limited to that specific user.
What are some suggestions for limiting access to the server to the user that started the application?
@GeorgeEdison On BSD and Linux systems there is a getpeereid() call to get the UID of whoever is at the other end of a socket; it is normally for Unix-domain sockets but it may work for TCP sockets which go through localhost (this is worth a try)
@GeorgeEdison It would imply the users typing their password again. It is always delicate to train your users into typing their passwords into random applications.
(these days, X11-over-network becomes rare because OpenGL-over-network does not cut it; but in older times, the .Xauthority file was quite common, and it worked)
besides the "use once" from @this.josh, you could also create a $HOME/foobar.html which is readable only by the user, and have the Web browser open that file
so, you open the URL with a nonce, then set a cookie with it and check it again for every subsequent request, from the cookie (which is easy if you replace your BasicHTTPServer with cherrypy)
@ThomasPornin While the U.S. Shuttle can descend in autopilot mode, normally the crew does take over the manual controls close to landing. John F. Hanaway and Robert W. Moorehead, Space Shuttle Avionics System (NASA SP-504, 1989), p. 25.
Interesting... I have a comment on this thread that hints at closure, and that comment has four up-votes, but the thread itself only has my single close vote.
The question is less about how to disclose information, and more about ways to ethically benefit from those disclosers. Unclear to me if the question is off-topic, but figured I'd ask anyway.
How can we design a market for information disclosure where individual security researchers may benefit economically in an ethical way?
Assume a market where the participants are governments, researchers (black and white), educational institutions, corporations, and non-profit non-governmental ...
Looks like you guys have it thoroughly covered. I know of one app that does something similar - sagemath.org. The code base is huge; it's a python re-write of mathematica/maple and the command notebook() launches a twisted server locally
they just use username / password combinations configured in the current user's settings, as far as I can see - I haven't looked that hard
I would like to know if it is save for the host system of a virtual machine (VM - VirtualBox OSE in my case) to execute malware.
Can a virus break out and read or write data from the host system? Can it establish an internet connection if I disable it in my VM?
Is a VM a safe environment to tr...
The question is less about how to disclose information, and more about ways to ethically benefit from those disclosers. Unclear to me if the question is off-topic, but figured I'd ask anyway.
