Am I being hacked?
in [General]
|
From: Yannick Mortier on 8 Apr 2009 11:06 2009/4/8 9el <lenin(a)phpxperts.net>: > On Wed, Apr 8, 2009 at 8:04 PM, Bob McConnell <rvm(a)cbord.com> wrote: > >> On Behalf Of Richard Heyes >> >> I set up a simple form to save comments on my webpage, and after just >> one >> >> day of going live, i'm getting weird comments up like this >> >> >> >> declare @q varchar(8000) select @q = >> >> 0x57414954464F522044454C4159202730303A30303A313027 exec(@q) >> >> >> >> >> >> I don't recognise this code - is this an attempt to do something >> nefarious, >> >> or nothing I should worry about? >> > >> > Looks like it may be. As long as you escape you SQL correctly using >> > mysql_real_escape_string() or the equivalent, you should be OK. >> >> Let me see if I got this right. The data you got from the form tries to >> set up a local variable, assigns it a hex string as a value, then tries >> to execute it. That definitely looks like an attempt to crack your >> server. It looks like the semi-colons were removed somewhere, so none of >> it actually runs. But you would probably need a set of dis-assemblers to >> find out what CPU that code was written for and what it actually does. >> >> Next question: You said there are multiple comments like this. How do >> they differ, if they do? Possibly they are trying code for different >> CPUs. >> >> Did you trace these back to the logs to see if they all come from one IP >> or subnet? Is there anywhere to report these attempts that would >> actually do any good, or should you just ban that IP. >> >> But this one goes into my journal as something to be prepared for. >> >> I think the danger these codes have should be discussed well. And how to > resist such attacks in your server and apps should also be discussed in > greater depth. > > regards > > Lenin > > www.twitter.com/nine_L > I just googled for that string. Seems like you are not the only victim. Sadly, I can't give you any more advice. -- Currently developing a browsergame... http://www.p-game.de Trade - Expand - Fight Follow me on twitter! http://twitter.com/moortier
From: Andrew Ballard on 8 Apr 2009 11:42 On Wed, Apr 8, 2009 at 10:04 AM, Bob McConnell <rvm(a)cbord.com> wrote: > On Behalf Of Richard Heyes >>> I set up a simple form to save comments on my webpage, and after just > one >>> day of going live, i'm getting weird comments up like this >>> >>> declare @q varchar(8000) select @q = >>> 0x57414954464F522044454C4159202730303A30303A313027 exec(@q) >>> >>> >>> I don't recognise this code - is this an attempt to do something > nefarious, >>> or nothing I should worry about? >> >> Looks like it may be. As long as you escape you SQL correctly using >> mysql_real_escape_string() or the equivalent, you should be OK. > > Let me see if I got this right. The data you got from the form tries to > set up a local variable, assigns it a hex string as a value, then tries > to execute it. That definitely looks like an attempt to crack your > server. It looks like the semi-colons were removed somewhere, so none of > it actually runs. But you would probably need a set of dis-assemblers to > find out what CPU that code was written for and what it actually does. > > Next question: You said there are multiple comments like this. How do > they differ, if they do? Possibly they are trying code for different > CPUs. > > Did you trace these back to the logs to see if they all come from one IP > or subnet? Is there anywhere to report these attempts that would > actually do any good, or should you just ban that IP. > > But this one goes into my journal as something to be prepared for. > > Bob McConnell > You don't need a disassembler; I already said what that string is intended to do. If it is allowed to run on Microsoft's SQL Server, the hex value is implicitly converted to the string "WAITFOR DELAY '00:00:10'", which is then executed. It doesn't require semi-colons, as SQL Server doesn't need them between statements. This particular command is relatively harmless by itself. Its value lies in the fact that if it causes the resulting page to take more than 10 seconds to load, the attacker knows that your page is wide open to SQL injection as well as knowing that he can execute anything he wants. If you're running MySQL, this won't work so you should be unaffected. Just make sure your code is written to prevent SQL injection and you should be fine. I suppose if you get a lot of these requests from the same IP address you could have the web server block requests from that IP. Andrew
From: "Warren Vail" on 8 Apr 2009 11:58 Looks like an attempt to get your SQL server to execute a command, Microsoft SQL server will do that(among others), and if not properly set up can do it with root access. If you don't properly escape and store this comment in a database, it could execute (called SQL injection, no?). Warren Vail > -----Original Message----- > From: Yannick Mortier [mailto:mvmortier(a)googlemail.com] > Sent: Wednesday, April 08, 2009 8:07 AM > To: 9el > Cc: Bob McConnell; Richard Heyes; julian haffegee; PHP Mailing List > Subject: Re: [PHP] Am I being hacked? > > 2009/4/8 9el <lenin(a)phpxperts.net>: > > On Wed, Apr 8, 2009 at 8:04 PM, Bob McConnell <rvm(a)cbord.com> wrote: > > > >> On Behalf Of Richard Heyes > >> >> I set up a simple form to save comments on my webpage, > and after > >> >> just > >> one > >> >> day of going live, i'm getting weird comments up like this > >> >> > >> >> declare @q varchar(8000) select @q = > >> >> 0x57414954464F522044454C4159202730303A30303A313027 exec(@q) > >> >> > >> >> > >> >> I don't recognise this code - is this an attempt to do something > >> nefarious, > >> >> or nothing I should worry about? > >> > > >> > Looks like it may be. As long as you escape you SQL > correctly using > >> > mysql_real_escape_string() or the equivalent, you should be OK. > >> > >> Let me see if I got this right. The data you got from the > form tries > >> to set up a local variable, assigns it a hex string as a > value, then > >> tries to execute it. That definitely looks like an attempt > to crack > >> your server. It looks like the semi-colons were removed > somewhere, so > >> none of it actually runs. But you would probably need a set of > >> dis-assemblers to find out what CPU that code was written > for and what it actually does. > >> > >> Next question: You said there are multiple comments like > this. How do > >> they differ, if they do? Possibly they are trying code for > different > >> CPUs. > >> > >> Did you trace these back to the logs to see if they all > come from one > >> IP or subnet? Is there anywhere to report these attempts > that would > >> actually do any good, or should you just ban that IP. > >> > >> But this one goes into my journal as something to be prepared for. > >> > >> I think the danger these codes have should be discussed > well. And how > >> to > > resist such attacks in your server and apps should also be > discussed > > in greater depth. > > > > regards > > > > Lenin > > > > www.twitter.com/nine_L > > > > > I just googled for that string. Seems like you are not the > only victim. Sadly, I can't give you any more advice. > > > -- > Currently developing a browsergame... > http://www.p-game.de > Trade - Expand - Fight > > Follow me on twitter! > http://twitter.com/moortier > > -- > PHP General Mailing List (http://www.php.net/) To > unsubscribe, visit: http://www.php.net/unsub.php >
From: "Michael A. Peters" on 8 Apr 2009 12:05 Andrew Ballard wrote: >> > > You don't need a disassembler; I already said what that string is > intended to do. If it is allowed to run on Microsoft's SQL Server, the > hex value is implicitly converted to the string "WAITFOR DELAY > '00:00:10'", which is then executed. It doesn't require semi-colons, > as SQL Server doesn't need them between statements. This particular > command is relatively harmless by itself. Its value lies in the fact > that if it causes the resulting page to take more than 10 seconds to > load, the attacker knows that your page is wide open to SQL injection > as well as knowing that he can execute anything he wants. If you're > running MySQL, this won't work so you should be unaffected. Just make > sure your code is written to prevent SQL injection and you should be > fine. Would doing all SQL queries via pear mdb2 prepare() and execute() prevent sql injection, or do other sql injection stepps need to take when using pear mdb2?
From: Andrew Ballard on 8 Apr 2009 12:51
On Wed, Apr 8, 2009 at 12:05 PM, Michael A. Peters <mpeters(a)mac.com> wrote: > Andrew Ballard wrote: > >>> >> >> You don't need a disassembler; I already said what that string is >> intended to do. If it is allowed to run on Microsoft's SQL Server, the >> hex value is implicitly converted to the string "WAITFOR DELAY >> '00:00:10'", which is then executed. It doesn't require semi-colons, >> as SQL Server doesn't need them between statements. This particular >> command is relatively harmless by itself. Its value lies in the fact >> that if it causes the resulting page to take more than 10 seconds to >> load, the attacker knows that your page is wide open to SQL injection >> as well as knowing that he can execute anything he wants. If you're >> running MySQL, this won't work so you should be unaffected. Just make >> sure your code is written to prevent SQL injection and you should be >> fine. > > Would doing all SQL queries via pear mdb2 prepare() and execute() prevent > sql injection, or do other sql injection stepps need to take when using pear > mdb2? > I'm not very familiar with the Pear libraries, but the approach is a good one. A good library will either pass the query off as a prepared statement with parameters to the database (if supported) or else will perform parameter substitution that should use mysql_real_escape_string() or similar functions to build a statement that should be safe to execute. The fact that you see these strings in the correct database field is a good sign, because they were correctly interpreted as string input and inserted rather than being interpreted as statements and executed. Just don't use this as a litmus test for your site's overall security. Andrew |