Flipper's Visual Basic 5 tough protection
(How to (try) to write a short app in Visual Basic and to protect it: a difficult task!)
tough

by Flipper (ucg)

(06 December 1997)


With some solutions:
The first solution, by r0ketman
and an answer by flipper
and another solution, by +Rcg
and another answer by flipper

Courtesy of reverser's page of reverse engineering
Well, I'm happy that Flipper decided to send a contribution to the "our protections" section. Hope many will tackle this target... Flipper is right: Visual Basic cracking may be pretty boring, but you learn a lot in the process! Yet read the letter by madmax! on our main our protections page... many believe that our protection schemes should be written in assembly!
Well, I agree with the assembly people: see, there is only on problem with Visual Basic 5 apps: the existence of Numega's Smartcheck! And using it onto this flipper's creation, you'll find something pretty interesting at event 25259! (MsgBox [VARIANT: String: "This Application will now terminate..."; and at event 84183: lbl_reg.caption: "Unregistered delay initiated..." and this is at APP1.EXE!000052F6, according to Smartcheck (I didn't dead list this little monster :-) and at event 84198 we have a timer(1) enabled (code:5350) and so on, and so on..;
Btw, flipper, according to Smartcheck your app has 2 errors, 12 leaks, 2246 VB and Windozes API and 18853 C/C++ and system API calls... ah yes, and 69 hooks as well :-)
I wrote some time ago a small essay darüber: An interesting tool: Numega's Smartcheck... I think that many people that will have a go trying to crack this flipper's target could find very useful help using this Numega's tool...

-- Overbloated Notepad v1.00 Documentation

-- A short summary of what this program is. 

Finally, an honest to goodness Visual Basic 5.0 Overbloated Notepad to call
your very own! What you could have written on a scrap piece of paper suddenly
comes to life in this wannabe MFC code, with no strings attached.

Download flipper's protection here (ucg10.zip, 15742 bytes) and work!

This notepad does not do anything other than to serve as a tool for learning
how to reverse enginner visual basic 5.0 programs. I myself tried a few
different approaches to patching this program, but each time the vbRuntime
DLL abomination stepped in and saved the day with another arcane error call.

Eventually I found my own methods of cracking the program, and figured that
since I could accomplish the job then _anyone_ could. That's where I began
to figure out new methods of stopping petty byte patches instead of real
keygens. What you have in front of you is probably one of the most advanced
visual basic protection schemes ever. That doesn't mean it is though, you'll
just have to try and crack it. The reason I say this, is that most vb
protections encountered by crackers can be defeated with a few set break-
points, whereas you may have to dead list this program to get a 'deeper' look
at it.

-- Registration Benefits.

If you register this program you'll get rid of that awfully slow nag, and
enable file saving (which is why you downloaded this program in the first
place, right?). Not to mention you'll gain valuable cracking skills in the
process 

-- How Do I Register?

Simple. Place œ10 in an old envelope.. and when you get your reg code back,
click on the "Register" option under the "About" menu and type in all your
misc. info.

-- Helpful hints for the beginner cracker.

Since this program was written in visual basic you can safely assume that
the code will be a mixture of function calls and random byte placings (as
with all Micro$oft programs). I never tried Soft-Ice out on this program
yet, so you're welcomed to do that. I also tried to keep the file size as
low as possible (it was more difficult that you think ) so anyone,
especially newbyes could use wdasm or IDA to poke around inside the multi-
volume listing.

 (a) read the configuration file carefully, and try to decipher the code
     in your head (okay, let's not go _that_ far..)

 (b) use softice and set breakpoints at vb sensitive calls and poke around
     the code a little.

 (c) I've used my own error handler routines, so if you decide to disable
     them, vb will jump in and save the day with it's own error handler.

 (d) There are no keyfiles involved in the protection, just good old
     calculations and the like (nothing a person with a calculator couldn't
     figure out..)

 (e) I won't reveal all of my protections, but rest assured nothing like
     this has ever before been attempted in visual basic. I don't mean the
     serial number routine, I mean the rest of the protection. Visual Basic
     programmers don't often know any assembler, and therefore have never
     tried out protections involving trig. functions, etc.

-- Acceptable Solutions

You can either write

 (i) a byte patch

     ; this will be the most commonly done crack by far, but you'll soon
       find out that it's not worth your time to try -- I've put up large
       detours and 'roadblocks' to handle generic byte patches. This should
       not discourage you though. If you find a way around all this, you're
       more than welcome to try.

 (ii) a keymaker

     ; if you can make one of these, even with all that messy vb code,
       then consider yourself a windows guru, it's just that impressive.

Of course, each should be accompanied by the method at which you arrived
to your final conclusion (ie: in essay format) and submitted to Reverser's
page at http://fravia.org/.

flipper (upg) 04/Dec/97

14 Dicember 1997: A first "answer-solution" by r0ketman!

Defeating Flipper's Visual Basic 5 tough protection scheme
The Easy Way.
---------------------------------------------------------
by r0ketman.
=========================================================

[tools you will need: Wdasm, Winice, and a hex editor.]


Intro.
======
This is my first submission, I hope everybody can understand it--I try
to keep things simple so you don't need an extra brain to figure out
what the heck I'm talking about. 

At first look, app1 appears to be a difficult crack. You say: "oh my
god! What is all of this vb tomfoolery??!?" Well I must admit, I too was
sucked into the despair mentality of "damn, I can't even crack a vb
app?" If you screw your head back on and use some common sense, you will
find it is not very difficult after all. In fact, this is as simple as
they get folks. 

A First Glance.
===============
You can spend all the time you want trying to figure out the damn algo
for the serial number--if spending time is your thing. I prefer to be
done as quickly as possible and that means in this case to byte patch.
Ok, on to the real info.

How to go about cracking app1.
==============================
Run app1 and see what it does. Do not rush to use tools without running
mr. exe to see what you are looking for. In the case of app1 we find
that giving a bad serial number will not display an error messagebox
whereas trying to save the file without registering the exe will. Ok,
now go into wdasm
or whatever you like and look at the imported functions. We find that
there is a function "rtcmsgbox."
Now, if we want to spend a million years tracing here and there in the
app we could set a bpx to rtcmsgbox, but lets look around some more in
the imported functions list. Hrmmm... Hey, wonder
what "rtcinputbox" does? It might be the function called to get the
serial number info. Load up winice 
and set a "bpx rtcinputbox" and run app1. If you click on the
registration menuitem it will not break which leaves us wondering what
rtcinputbox does. Considering the amount of options available in the
program, it would be a likely assumption that it is in fact the dialog
entry for saving files. Lets go back into wdasm and search for
rtcinputbox.

We get one hit:

* Ref to MSVBVM50.rtcInputBox
:00403F80 Call 00401366

Well uhh what else could that go to other than a file entry dialog? Lets
trace up...
Aha!

:00403E85 mov ax, word ptr [00408030]
:00403E8B xor esi, esi
:00403E8D cmp ax, 03FB
[lots of mov instructions here]
:00403EC4 jnl 00404035
:00403ECA cmp ax, 001B
:00403ECE jle 00404035
:00403ED4 cmp ax, 00FE
:00403ED8 jle 00404030

Well, we dont want to jump to any of those locations so the first number
that comes to mind that will keep it from jumping would be ax=00FF. Ok
thats logical enough. We can give it a whirl in winice:

bpx 00403E8D
>break due to BPX 0137:00403E8D
r ax=00FF
cntl+d

Goody goody, it works! We now get a file save dialog and it appears to
actually save the file. But that still leaves the nag to get rid of.
Well at this point I was assuming that 00FF was the "goodguy" flag and I
got inspired to go back to wdasm and search for 00FF hoping for an
obvious mov function.

Wdasm hit on:

:0040129F add bh,bh   [no i dont think this is what we want]
:004016D8 add bh,bh   [nope]
...
:00405242 mov word ptr [00408030], 00FF [finally! hey, loc 408030 is familiar!]

Ok, I think we found our bad boy. I don't know much about visual basic,
but I don't think its a coincidence that right before the mov there is a
reference to MSVBVM50.__vbaVarTstEq. Guess that stands for
VisualBasic_Test_Equation--I dunno. Well, the whole snippet looks like:

* Ref to MSVBVM50.__vbaVarTstEq
:00405232 Call 0040143E
:00405237 test ax, ax			  [ax=0000 no wonder it jumps all the time]
:0040523A je 00405293                     [hey if we NOP this we can get to the badboy]
:0040523C cmp dword ptr [00408010], edi
:00405242 mov word ptr [00408030], 00FF   [our badboy, we must not be hitting this loc] 
:0040524b jne 0040525C			  [leave this alone or we will wind up in a loop]

Now Flipper states that his program will check if its been tampered
with. Err, lets suppose I'm stupid and I missed that part. Using Joe
Blow Hex Editor, change 7457h at offset 463Ah [:0040523A je 00405293] to
9090h. Ok, the badboy has been NOP'd. Lets run app1 to see if it worked.
Ok, no nag. Put some junk
text in and click save. Hey!, it seems to be fully operational without
actually registering the program.

The End.
========

Here we are after a few minutes at the end of my text. See, it wasn't
that hard now was it? This [app1] is certainly an interesting try at
visual basic protection, but I doubt any amount of trig func's are going
to keep people from finding an easy way out. Visual
Basic has a nasty habit of naming functions that we can breakpoint on
that are very difficult to avoid programming-wise. In sum,
when dealing with Visual Basic apps, try to think simple as the more
complex a method you try the use, the more likely it will
be more difficult to crack.

the r0ket.
wicketimp@hotmail.com


14 Dicember 1997: A first answer by flipper!
  That was indeed a very well written essay with a sound approach. I'd like
to say "ya got me" to rOcket, as I did not anticipate him byte patching that
particular function. Some explaination is required of course to why my
program didn't snap when he changed that byte :-)

  I didn't have time to write a full blown CRC check on the whole EXE, so I
rigged up a makeshift byte checker. The file APP1.DAT contains offsets (in
decimal), and correct byte patterns for those offsets xor'd of course. The
formula to 'decrypt' the values is not important, but the truth is, I had
only placed checks on what I thought would scare newbies away at first
glance 

  The byte checks are at the comparison routines jge, jbe, etc. where I
thought most new crackers would try and patch the program. Also, there's a
check around the timer call to stop anyone from changing that.

  What would have worked much better was a full EXE check, but even on a
pII-300 it would take eons using visual basic! I have also found a nice
trick to pin down any message box in VB to the dead listed code.

  You may see a 'push 00401234' and get no string reference to that, so I
have a little trick for soft-ice and vb apps. I think someone else used this
before, but not quite in this way. Also, with this method, there's no need
to understand the call relocation table (but you should!).

  Load the target app, wait for a messagebox or event, etc. Set your
breakpoints, and when soft-ice snaps, press F12 (assuming it is still 'p
ret') as many times as it takes to get back to the application code [watch
the green status bar between the command window and code window]. Once
there, you have the dead listing offset where your function lies, say at
014f:00405211. Switch to IDA or wdasm32, and have a look. You'll see
something like 'call [ebx+08]' maybe, and that's the call to the messagebox
or whatever. This approach also works with VC apps, and just about anything.

  Of course, you'll have to still 'feel' the code and decide where to patch,
but I'll leave that up to the cracker.

"I have been defeated, but will be back with a visual basic app that'll do
away  with smartcheck and perhaps softice.." ;-)

Best Regards,

  flipper (upg)

14 Dicember 1997: A second "answer-solution" by +Rcg!
How to reverse Overbloated Notepad by Flipper (upg).

After all as usual lets 'enter in contact' with the program and then
extract some conclutions:

	* Register option is available.
	* Heavy disk activity.
	* Check for a good 'Code' is done at next sesion.
	* App1.dat is necesary.
	* App1.cfg is generated if it doens't exist.
	* It was writen in VB :-)
	* A nagscreen appears if you are 'Unregistered' .-)
	 

Why there are so much disk activity?
Let's use FileMonitor filtered with App1 and the current directory.

Seek	\APP1.EXE	SUCCESS	 Beginning Offset: 1024	
Read	\APP1.EXE	SUCCESS	 Offset: 1024 Length: 4096	
Seek	\APP1.EXE	SUCCESS	 Beginning Offset: 5120	
Read	\APP1.EXE	SUCCESS	 Offset: 5120 Length: 4096	
Seek	\APP1.EXE	SUCCESS	 Beginning Offset: 28160	
Read	\APP1.EXE	SUCCESS	 Offset: 28160 Length: 512	
Seek	\APP1.EXE	SUCCESS	 Beginning Offset: 9216	
Read	\APP1.EXE	SUCCESS	 Offset: 9216 Length: 4096	
Seek	\APP1.EXE	SUCCESS	 Beginning Offset: 21504	
Read	\APP1.EXE	SUCCESS	 Offset: 21504 Length: 4096	
Open	\APP1.EXE	SUCCESS	 OPENEXISTING OPENALWAYS 	
Open	\APP1.EX~	NOTFOUND 	 CREATEALWAYS REPLACEEXISTING 	
Open	\APP1.EX~	SUCCESS	 CREATENEW CREATEALWAYS 
				        OPENALWAYS REPLACEEXISTING 	
Read	\APP1.EXE	SUCCESS	 Offset: 0 Length: 65024	
Write	\APP1.EX~	SUCCESS	 Offset: 0 Length: 35840	
Read	\APP1.EXE	SUCCESS	 Offset: 35840 Length: 65024	
Close	\APP1.EX~	SUCCESS		
Close	\APP1.EXE	SUCCESS		
Open	\APP1.EXE	SUCCESS	OPENEXISTING OPENALWAYS 	
Open	\APP1.EX~	SUCCESS	OPENEXISTING OPENALWAYS 	
Attrib.\APP1.EXE	SUCCESS	Get Creation	
Attrib.\APP1.EXE	SUCCESS	Get Access	
Attrib.\APP1.EXE	SUCCESS	Get Modify	
Attrib.\APP1.EX~	SUCCESS	Set Creation	
Attrib.\APP1.EX~	SUCCESS	Set Access	
Attrib.\APP1.EX~	SUCCESS	Set Modify	
Close	\APP1.EXE	SUCCESS		
Close	\APP1.EX~	SUCCESS		
Open	\APP1.EX~	SUCCESS	CREATENEW CREATEALWAYS 
					OPENEXISTING OPENALWAYS 	
Open	\APP1.DAT	SUCCESS	OPENEXISTING OPENALWAYS 	
Read	\APP1.DAT	SUCCESS	Offset: 0 Length: 512	
Seek	\APP1.EXE	SUCCESS	Beginning Offset: 13312	
Read	\APP1.EXE	SUCCESS	Offset: 13312 Length: 4096	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 12941	
Read	\APP1.EX~	SUCCESS	Offset: 12941 Length: 5	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 12996	
Read	\APP1.EX~	SUCCESS	Offset: 12996 Length: 28	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 13992	
Read	\APP1.EX~	SUCCESS	Offset: 13992 Length: 5	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 14014	
Read	\APP1.EX~	SUCCESS	Offset: 14014 Length: 22	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 17626	
Read	\APP1.EX~	SUCCESS	Offset: 17626 Length: 8	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 18313	
Read	\APP1.EX~	SUCCESS	Offset: 18313 Length: 5	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 23555	
Read	\APP1.EX~	SUCCESS	Offset: 23555 Length: 9	
Read	\APP1.DAT	SUCCESS	Offset: 220 Length: 512	
Seek	\APP1.EXE	SUCCESS	Beginning Offset: 25600	
Read	\APP1.EXE	SUCCESS	Offset: 25600 Length: 2560	
Close	\APP1.DAT	SUCCESS		
Close	\APP1.EX~	SUCCESS		
FindOp.\APP1.EX~	SUCCESS	APP1.ex~	
FindOp.\APP1.EX~	SUCCESS	APP1.ex~	
FindCl.\APP1.EX~	SUCCESS		
Delete	\APP1.EX~	SUCCESS		
FindNe.\APP1.EX~	NOMORE		
FindCl.\APP1.EX~	SUCCESS		
Seek	\APP1.EXE	SUCCESS	Beginning Offset: 17408	
Read	\APP1.EXE	SUCCESS	Offset: 17408 Length: 4096	
Open	\VALIDATE.DAT	NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\APP1.KEY	NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\CHCKSUM.DAT	NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\A_VERIFY.REG	NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\EB6A		NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\EB6B		NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\EB6C		NOTFOUND OPENEXISTING OPENALWAYS 	
..
..
Open	\ED9H		NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\ED9I		NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\ED9J		NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\PATCH.EXE	NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\A.DAT		NOTFOUND OPENEXISTING OPENALWAYS 	
Open	\APP1.CFG	SUCCESS	 OPENEXISTING OPENALWAYS 	
Read	\APP1.CFG	SUCCESS	 Offset: 0 Length: 512	
Read	\APP1.CFG	SUCCESS	 Offset: 180 Length: 512	
Close	\APP1.CFG	SUCCESS		



Conclutions:

	1. Files named EB6A,EB6B...ED9J  are generated
	   by a bucle and are fog to our eyes...so ignore them.
	2. Files APP1.KEY, VALIDATE.DAT, CHCKSUM.DAT and
	   A_VERIFY.REG are only opened and closed (not read),
	   so it may only check for it existence.
	3. The only file read is APP1.DAT.
	4. File App1.ex~ is copy of App1.exe ==> file check
						 integrity?

Let's use the normal aproach to all UserName/RegisterCode schemes.
bpx at nagscreen====>bpx rtcmsgbox (this is the MessageBoxA 
VB function) and F11.

Now, write down the address (4055A0) and W32Dasm.


* Referenced by a CALL at Address:00404030   
|
405512     push ebp
405513     mov ebp, esp
..
..
405595     lea eax, dword ptr [ebp-20]
405598     push 00000010
40559A     push eax
40559B     Call 00401396		;rtcMsgBox
4055A0     lea eax, dword ptr [ebp-50]


And at 404030:

* Referenced by a  Jump at Address:403ED8(C)
|
404030 E8DD140000              call 00405512


So 

* Referenced by a Jump at Address:0040252A(U)
|
403E43    push ebp
403E44    mov ebp, esp
403E46    sub esp, 0000000C
403E49    push 004012A6
403E4E    mov eax, dword ptr fs:[00000000]
403E54    push eax
403E55    mov dword ptr fs:[00000000], esp
403E5C    sub esp, 000000F8
403E62    mov eax, dword ptr [ebp+08]
403E65    and dword ptr [ebp+08], FFFFFFFE
403E69    and eax, 00000001
403E6C    mov [ebp-08], 00401038
403E73    push ebx
403E74    mov dword ptr [ebp-04], eax
403E77    mov eax, dword ptr [ebp+08]
403E7A    push esi
403E7B    push edi
403E7C    mov ebx, dword ptr [eax]
403E7E    mov dword ptr [ebp-0C], esp
403E81    push eax
403E82    call [ebx+04]
403E85    mov ax, word ptr [00408030] <== Ok this is our flag!!! 403E8B xor esi, esi 403E8D cmp ax, 03FB 403E91 mov dword ptr [ebp-18], esi .. .. 403EB8 mov dword ptr [ebp+FFFFFF50], esi 403EBE mov dword ptr [ebp+FFFFFF40], esi 403EC4 jnl 00404035 403ECA cmp ax, 001B 403ECE jle 00404035 403ED4 cmp ax, 00FE 403ED8 jle 00404030 <="=" Here .. So the flag is at [408030]: But, where is modified our flag (searching for 408030 we find ocurrences at: 403E85 4042A0 404B95 405242 405296 Let's see: :00403E81 50 push eax :00403E82 FF5304 call [ebx+04] :00403E85 66A130804000 mov ax, word ptr [00408030] :00403E8B 33F6 xor esi, esi :00403E8D 663DFB03 cmp ax, 03FB :0040429C 50 push eax :0040429D FF5104 call [ecx+04] :004042A0 66A130804000 mov ax, word ptr [00408030] :004042A6 33FF xor edi, edi :004042A8 663DFB03 cmp ax, 03FB :00404B8E 5E pop esi :00404B8F 8D8D18FFFFFF lea ecx, dword ptr [ebp+FFFFFF18] :00404B95 66893D30804000 mov word ptr [00408030], di :00404B9C 89BDD4FEFFFF mov dword ptr [ebp+FFFFFED4], edi :00404BA2 89B5CCFEFFFF mov dword ptr [ebp+FFFFFECC], esi :00405231 50 push eax :00405232 E807C2FFFF Call 0040143E ;__vbaVarTstEq :00405237 6685C0 test ax, ax :0040523A 7457 je 00405293 :0040523C 393D10804000 cmp dword ptr [00408010], edi :00405242 66C70530804000FF00 mov word ptr [00408030], 00FF :0040524B 750F jne 0040525C :0040524D 6810804000 push 00408010 :00405293 8B4508 mov eax, dword ptr [ebp+08] :00405296 66C70530804000FE00 mov word ptr [00408030], 00FE Ok!!! solution is in front of you at this time!!!!! Look at 405237="="> FF Good Guy.
		   FE Bad  Guy.

bpx at 405232 and wait.
Look at the stack:

ss:esp+0 ==> 2nd Value to Tst
ss:esp+4 ==> 1st Value to Tst

e ss:esp+0 => arguments to the function, looking at address 
referenced by the 3er dword  you can see the correct code.

So: Name: +HCU
    Code: 6AB76748

Let's see the posible file integrity check.

Seek	\APP1.EX~	SUCCESS	Beginning Offset: 12941	
Read	\APP1.EX~	SUCCESS	Offset: 12941 Length: 5	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 12996	
Read	\APP1.EX~	SUCCESS	Offset: 12996 Length: 28	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 13992	
Read	\APP1.EX~	SUCCESS	Offset: 13992 Length: 5	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 14014	
Read	\APP1.EX~	SUCCESS	Offset: 14014 Length: 22	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 17626	
Read	\APP1.EX~	SUCCESS	Offset: 17626 Length: 8	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 18313	
Read	\APP1.EX~	SUCCESS	Offset: 18313 Length: 5	
Seek	\APP1.EX~	SUCCESS	Beginning Offset: 23555	
Read	\APP1.EX~	SUCCESS	Offset: 23555 Length: 9


And now look at app1.dat

14989 6338FE068C
15044 0A886E04050563381E050A8B640405056338FB050A8B570405056F0F
16040 6338FE068C
16062 0A888E05050563381E050A8B840505056338FA05787E
19674 3EC20A81A9070505
20361 EC49F8FAFA
25603 5755FA545D80C5780B

Do you see something in common????????

14989-2048=12941
15044-2048=12996
16040-2048=13992
16062-2048=14014
19674-2048=19626
20361-2048=18313
25603-2048=23555


12941: 66 3D FB 03 89  (from App1.exe offset 12941) 
Xor:   63 38 FE 06 8C  (from App1.dat)
--------------------- 
       05 05 05 05 05 

12996: 0F 8D 6B 01-00 00 66 3D 1B 00 0F 8E 61 01 00 00 66 3D FE ...
Xor:   0A 88 6E 04 05 05 63 38 1E 05 0A 8B 64 04 05 05 63 38 FB ...
-------------------------------------------------------------------
       05 05  05 05 ...


and more....


Ok!!! now you can patch is because you know the correct code in
file App1.dat.


But, why it not checks at:

:00405231 50                      push eax
:00405232 E807C2FFFF              Call 0040143E  ;__vbaVarTstEq
:00405237 6685C0                  test ax, ax
:0040523A 7457                    je 00405293 <== Here 'Nopping' at 40523A is more than enough to patch the program. The 'KeyMaker' Let's use a zen approach. Let's make a table using different UserNames and getting the correct code. UserName Code "A.class" tppabs="http://fravia.org/A.class" 6AB65BEC AA 6AB6B1F4 AAA 6AB707FC B 6AB65D3F BB 6AB6B499 BBB 6AB70BF4 C 6AB65E91 CC 6AB6B73F CCC 6AB70FED . . . Z 6AB67D02 ZZ ... ZZZ ... + 6AB63ECD Nothing 6AB605E3 Look, the increment between two consecutive values can be 152h or 153h. Lets make the next operation: value of nothing-value of letter. Letter Value at first position difference at other positions + 38EA EVEN ODD A 5609 (-1) (+1) B 575B (-1) (0) C 58AD (+1) (0) D 5A01 () E 5B54 () F 5CA7 () G 5DF9 (0) (0) H 5F4C (0) (O) I 609F () J 61F2 () K 6345 () L 6498 (-1) (0) M 65EA () N 674C () O 6890 () P 69E3 (0) (-1) Q 6B36 () R 6C89 (+1) (0) S 6DDB () T 6F2D () U 7081 (0) (0) V 71D4 () W 7327 () X 747A () Y 75CC () Z 771E () The formula is="=">  6AB605E3+(sum of every letter)+differences

Now, lets create some codes:

	+ABC = 6AB605E3+(38EA+5609+575B+58AD)-1+1=6AB744DE
	+HCU = 6AB605E3+(38EA+5F4E+58AF+7085)+0=6AB76748
	+ALL = 6AB605E3+(38EA+5609+6498+6498)-1-1=6AB75E04
	+RCG = 6AB605E3+(38EA+6C89+58AD+5DF9)+1+0+0=6AB761FD
	APP =  6AB605E3+(5609+69E3+69E3)+0-1=6AB72FB1
	

Now, you can find out the missing values and create your own
keymaker.


+Rcg 1997


26 Dicember 1997: A second answer by flipper!
Reverser+,

  I'm sorry for the long delay, I've been busy as of late, but here's a
response to +RCG's essay on my Visual Basic protection.

  For the second time now, I've seen my Visual Basic protection 'byte the
dust'. It seems that +RCG has a certain routine he follows whenever he cracks
an app. Using filemon is an excellent way of seeing which files are fetched
by the program. I set up my program to create random files galore, but it
didn't fool +RGC (a seasoned +cracker).

  It just goes to show how insecure VB really is. You can use SoftIce,
SmartCheck, and a dead list (works quite well for vb5 apps) among other
things to crack any vb program. Since my protection has been defeated twice
now :-) the full source code will be published as well so that anyone who
chooses to compare dead lists with actual source code can keep track of the
soup mix we've all come to known as Windows.

  The reason I didn't use an external DLL to perform a CRC check on the file
is (1) an OCX was not available, and (2) would be easily defeated by anyone
who has read "DLL Based Protections Are Dead" from your page. So I made my
first fatal mistake; copying the actual program to a temp file and then
performing the necessary byte checks on that copy instead of the original
(VB can't bloody well do that either). Filemon was able to reveal exactly
which bytes I was reading, so RCG was able to figure out quite easily that
the locations in APP1.DAT were only 2048 bytes away fron their actual
position. He also figured out (I won't ask how long this took!) that I XOR'd
each byte by 05 to scare off the newbie crackers. I only wish I could have
implemented a CRC32 check (at the least) to thwart off our good friend +RCG.

  That about covers it, and as I learn more about cracking, I'll apply those
tips and techniques to try and develop an application that is next to
impossible to crack. It's a dream, I know.. but I already have a few plans
in the works.

  My next protection will be in VB 3.0 (I'll be using Visual C later on in
February), so it's going to be 16bit. Also, it will not decompile very
easily and it will kick SmartCheck, BoundsChecker, FileMon, OpenTrap, etc.
out of memory before it starts up. The last thing I'll try to implement is
some nice anti soft-ice code that will hopefully kick that code collector
out of memory also. You'll just have to wait and see. Hopefully I can
deliver on these outrageous promises ;) Until next time.

Best Regards,

  flipper (upg) 12/25/1997

(c) flipper (upg) All rights reversed
You are deep inside reverser's page of reverse engineering, choose your way out:

redBack to 'Our Protections'
redhomepage redlinks redanonymity +ORC redstudents' essays redacademy database
redtools redcocktails redantismut CGI-scripts redsearch_forms redmail_fravia
redIs reverse engineering legal?