5 ms·
BackTrack successor Kali Linux launched
- rman666 14y agoWhere's the distro? It doesn't seem to be up at http://www.backtrack-linux.org http://www.backtrack-linux.org yet.
- rman666 14y agoAh, ha! http://www.kali.org/ http://www.kali.org/
- zozu 14y agoTime to slap in in VirtualBox tonight and start to tinker with it.
- wuest 14y agoLooks pretty sweet! I'll be digging in tonight. Looks like there's even an official build for the ODroid U2! Sweet deal, I happen to own two.
- q3k 14y agoTo be honest, I still don't see the need for such a distribution to exist, as all the packages I ever needed during pentesting were available on Gentoo's Portage. When I've tested it, the mental overhead of having to run it on a VM, or even dual-boot it was quite annoying.
- EthanHeilman 14y agoIt was my understanding that you often need kernel patches to sniff traffic over ethernet or wifi. Backtrack solves this problem by having a custom kernel such that if you buy laptop with chipset A you can sniff traffic without serious work. My knowledge on this stuff is a little old, is this not the case anymore?
- shocks 14y agoIt depends on your WiFi card. Most Intel cards are pretty much plug and play in terms of packet injection and wireless cracking, but some strange WiFi cards can be hard to get working.
- q3k 14y agoNot really much of a problem with the chipsets that I used (ralink ones) - the stock kernel drivers work fine. But I do remember that there used to be some problems with that. The fact that I use a distro which allows me to painlessly integrate a custom kernel into my system also helps - any sane patches could have been applied without any problem.
- lawnchair_larry 14y agoConvenience. Portage is quite a waste of billable consulting hours. You do have to change the artwork though, because these backtrack/Kali distros and their slogans have rather silly/childish themes.
- q3k 14y agoMaybe when you're not prepared - I used to have all the tools necessary emerged and ready to use. Not to mention that most tools that I met compiled under five minutes.
- ErikRogneby 14y agoWith the ARM/Raspberry PI support I fully expect to see this get productized in to a low cost "pentest in a box".
- amckenna 14y agoyeah it would be great to see a low cost version of the products offered by pwnie express such as http://pwnieexpress.com/products/pwnplug-wireless http://pwnieexpress.com/products/pwnplug-wireless
- runjake 14y agoThat is OffSec's goal with this. They were on a recent infosec podcast and specifically mentioned providing a low cost alternative to the PwnPlug using hardware like the RaspPi.
- ilconsigliere 14y agoIs pen testing as fun as it seems like it'd be? I want to get into it but seems like a dangerous pool to jump into on your own.
- freehunter 14y agoIt's dangerous if you start running against production sites, or sites that aren't fully owned by you. If you broke into Microsoft.com, for example, expect a lawsuit. However, if you pair two machines together and run Backtrack/Kali on one and something along the lines of Damn Vulnerable Linux on the other and just attacked your own local network, it's fun, safe, and informative. I would actually encourage developers to learn about pen testing. If you know how people are going to misuse your application, you know what to watch out for when you're designing it. And trying to break your own app gives you some new insight into what you're doing right and what you're doing wrong. You can feel a sense of pride and accomplishment for every attack that fails to break something.
- wuest 14y agoShort answer: yes. Long answer: Sometimes. It's also one of the most frustrating things you can engage in. Up until the point where you have a 'eureka' moment, you feel like every step you've taken was through a tar pit. Pentesting is definitely full of pain; it's easy to imagine it's nothing but an adrenaline rush, since that's what people are inclined to discuss--but that's not the case at all. Edit to answer the other half of the statement: Pentesting is not dangerous. Malicious network activity (i.e. network activity which the target does not desire) is dangerous. There are lots of great resources (OffSec offers fantastic training, for example) on the topic, and there are gobs of free resources as well (CoreLan, interactive stuff like SmashTheStack's CTFs, DamnVulnerable Linux, etc.) for learning about the topic--tools, methods and such.
- runjake 14y agoParts of it can be fun, but much of the time is spent writing tedious (and repetitive) reports, documenting each step you took for each compromise. Dealing with unrealistically limited scopes. Often clients don't actually care about increasing their security posture -- it's usually about compliance with XYZ (PCI, etc). (Good) pentesters don't just run hog wild on a network compromising things left and right with Chemical Brothers blaring through their headphones. It's pretty methodical, and it's all documented and screenshoted for the client's benefit. You want your clients, both the technical and executive people in the org, to understand what you accomplished and how they can mitigate.
- faster 14y agoGrabbed the torrent. The md5sum of the ISO doesn't match the one listed in the included txt file. This concerns me a little.
- eruditely 14y agoNot enough attention given to this reply.
- tripzilch 14y agodid the same, also got a mismatch. then I realized, that the hash in the txtfile is also longer than the md5 sum, and facepalmed a little. with 40 hexdigits long, makes 160 bits, so it's probably SHA1. (crunch, crunch) ... and it matches. would've been nice if they'd mentioned which algorithm they used so I didn't have to digest a 2GB file twice. we can't be the only people that don't immediately eyeball the difference between a 32 and 40 digit random hex code :)