Scylla x64/x86 Imports Reconstruction

[Carbon]

Version 0.9.4 Final

- direct import scanner (LEA, MOV, PUSH, CALL, JMP) + fixer with 2 fix methods
- create new iat in section
- fixed various bugs

I really recommend to update due to the bug fixes.

Direct import scanner fix methods:
- Normal: Patch memory with jmp/call only
- Universal: Works with everything, creates a jump table in the scylla section, watch for relocation information in the log file

I also found some weird thing in Windows 7 x64. I don't know yet why this happens:

### Windows 7 x64

Sometimes the API kernel32.dll GetProcAddress cannot be resolved, because the IAT has an entry from apphelp.dll
Solution? I don't know

[ahmadmansoor]

Originally Posted by Carbon

I really recommend to update due to the bug fixes.

Direct import scanner fix methods:
- Normal: Patch memory with jmp/call only
- Universal: Works with everything, creates a jump table in the scylla section, watch for relocation information in the log file
:

I was watch ur update ,My friend Universal import scanner fix is a Good Idea .
but it is limited with some Protector ,in other it is Difficult to handle it .
Let take the Themida/Winlicense : through the unpacked rutine ,it pass through IAT Table rebuild which write the API to the file .here it decide to write the

NOP
Jmp xxxxx
or
Call xxxxx
Nop

so this nop it Defined through this rutine ,and I think it is random .

00412893 CC int3
00412894 > 90 nop
00412895 .- E9 96287477 jmp msvcr100.__set_app_type
0041289A > 90 nop
0041289B .- E9 60587477 jmp msvcr100._amsg_exit
004128A0 > 90 nop
004128A1 .- E9 3A647477 jmp msvcr100.__wgetmainargs
004128A6 CC int3
+++++++++++++++++++++++++++++++++++++
004129C7 CC int3
004129C8 > 90 nop
004129C9 .- E9 D2567477 jmp msvcr100._exit
004129CE > 90 nop
004129CF .- E9 BCA68177 jmp msvcr100._XcptFilter
004129D4 >- E9 E7567477 jmp msvcr100._cexit
004129D9 . 6F outs dx, dword ptr es:[edi]
004129DA >- E9 A1567477 jmp msvcr100.exit
004129DF 13 db 13
004129E0 > 90 nop
004129E1 .- E9 DA708177 jmp msvcr100._CrtSetCheckCount
004129E6 CC int3

so guessing which NOP is the right to replce for Fix This import will fault by 70%

pls check this Image :
http://postimg.org/image/6fzu4kr8v/
and u will see what I was talking about .I have write a lot of tut on rebuild IAT for Themedi I can send it to u and through this tut u will see when and where the nop is written .
and so on for other Protector ,which each one his privacy .

I also found some weird thing in Windows 7 x64. I don't know yet why this happens

can u give example (code or File ) ?

Thanks for ur great work ,pls keep up.

[Carbon]

@giv
feel free to report bugs.

@ahmadmansoor
Try the "universal" direct import fixer (enable in options). It will work with Themida and any other protector.

I don't think I can give an example. It is still weird. It has probably something to do with this https://forum.tuts4you.com/topic/34548-scylla-version-announcements/#entry159332

[ahmadmansoor]

Originally Posted by Carbon

@ahmadmansoor
Try the "universal" direct import fixer (enable in options). It will work with Themida and any other protector.

my friend the example which I gave u in the Picture was universal enable in options I will upload the files when back to home .

I don't think I can give an example. It is still weird. It has probably something to do with this https://forum.tuts4you.com/topic/34548-scylla-version-announcements/#entry159332

I will check this

[Carbon]

Originally Posted by ahmadmansoor

my friend the example which I gave u in the Picture was universal enable in options I will upload the files when back to home .

Now I see there is a bug. You must disable the "normal" fixer otherwise the "universal" will not work. And it is fixed only in the dumped and fixed file. Not in memory.

[Computer_Angel]

Originally Posted by Carbon

I also found some weird thing in Windows 7 x64. I don't know yet why this happens:

We could using plugin for apphelp.dll to solve the api. This is my small plugin for Imprec & Scylla.

About scylla crash, I had found that the function ApiReader:arseExportTable is parsing export not correct in some case, the way of calculating functionName = (char*)(addressOfNamesArray[i] + deltaAddress) is not right if the address of names in the differ memory than the exportbuffer cover.