Difference between revisions of "Reverse-Engineering"
m (→Androguard) |
m (→Android) |
||
Line 302: | Line 302: | ||
===GameCIH=== |
===GameCIH=== |
||
===GameGuardian=== |
===GameGuardian=== |
||
+ | ===[https://labs.mwrinfosecurity.com/tools/drozer/ Drozer]=== |
||
==Java== |
==Java== |
Revision as of 19:23, 15 October 2013
Books
- The IDA Pro Book, 2nd Edition by Chris Eagle
- Reverse Engineering Code with IDA Pro by Dan Kaminsky et al
- Practical Malware Analysis by Michael Sikorski
- Reversing: Secrets of Reverse Engineering by Eldad Eilam
- Crackproof Your Software by Pavol Cerven
- Surreptitious Software: Obfuscation, Watermarking, and Tamperproofing for Software Protection
- Wikibooks Subject:Software_reverse_engineering
Resources
- Reverse-Engineering on StackExchange
- OpenRCE
- Hex Blog
- http://www.reverse-engineering.info
- Automating RE with Python (slides) by Carlos Prado
- Intel® 64 and IA-32 Architectures Software Developer Manuals
Static Analysis Tools
IDA Pro
IDA Pro combines an interactive, programmable, multi-processor disassembler coupled to a local and remote debugger and augmented by a complete plugin programming environment.
- Official page
- Windows, Linux, Mac OS X
- x86-32, x86-64, ARM and many others
- ELF, Java bytecode, Dalvik, ARM,...
- disassembler, some debugger
- extensible through plugins & python (anti-debugger, findcrypt,...)
- IDA toolbag
- IDAscope
- patchdiff2
- Zynamics bindiff
- DarunGrim, another binary diff tool, opensource but discontinued?
- x86emu, x86 Emulator plugin. Windows, Linux, OS X
- Plugin contests 2012, 2011, 2010, 2009
Hex-Rays
The most expensivepowerful IDA Pro plugin is the Hex-Rays decompiler
- x86 and ARM
- decompiler
Limitations specific to ARM:
- floating point instructions are not supported
- VFP/SIMD/Neon/... instructions are not supported
- functions having an argument that is passed partially on registers and partially on the stack are not supported (e.g. int64 passed in R3 and on the stack)
REC Studio
- x86, x64
- Windows, Linux, Mac OS X
- HLA disassembler
Useful commands:
help strings calltree showprocs decompile /tmp/myprog.c
click on a function in the "Project" function list to HLA disass it
Radare
The reverse engineering framework
Misc
Distorm
diStorm3 is really a decomposer, which means it takes an instruction and returns a binary structure which describes it rather than static text, this is great for advanced binary code analysis
PyPEELF
PyPEELF is a multi-platform binary editor written in Python, wxPython and BOA Constructor. It allows you to manage binary data in PE32, PE32+ (x64) and ELF binary files.
PyPEELF uses pefile to manage PE32 and PE32+ files and pyelf to manage ELF files. Besides, it uses winappdbg and pydasm in some others features like Task Running Viewer and Disassembling files.
PyPEELF was designed for Reverse Engineers who want to edit or visualize binary file data in multi-platforms. That is why PyPEELF runs under Windows and Unix/BSD operating systems
Poor man's tools
File, -z to uncompress, -s to inspect non-files, e.g. /dev/sda1
file -k [-z] [-s] mybin
Strings
strings [-n min_length] -a -e [s|S|b|l|B|L] mybin
Android
Documentation
- Dalvik: bytecode, dex & VM instructions
Dex2jar
A tool for converting Android’s .dex format to Java’s .class format
See also DeObfuscate jar with dex tool
./d2j-dex2jar.sh myapp.apk
This returns a file myapp-dex2jar.jar
Then use Java decompilers: jad, jd-gui, cf below
Smali
smali/baksmali is an assembler/disassembler for the dex format used by dalvik, Android’s Java VM implementation
Apktool
It is a tool for reverse engineering 3rd party, closed, binary Android apps. It can decode resources to nearly original form and rebuild them after making some modifications; it makes possible to debug smali code step by step. Also it makes working with app easier because of project-like files structure and automation of some repetitive tasks like building apk, etc.
apktool d myapp.apk
Apk Multi-Tool
Swiss knive (was Apk Manager)
Contains apktool, smali/baksmali etc
on Github for Linux release
- 9 decompile apk / 1 select apk / 9 decompile apk
GetStrings
Small script to prepare a sed script to inject resource strings into jad, to ease reversing.
To be used e.g. after apktool / Apk Multi-Tool decompilation
#!/bin/bash
DECOMPILED_DIR=working/*apk/
cat $DECOMPILED_DIR/res/values/public.xml|grep "type=.string"|\
sed 's/.*name="\?//;s/" id="\?/ /;s/"\? \/>//'|\
awk --non-decimal-data '{print $2, int($2), $1}'\
> getstring-pub
cat $DECOMPILED_DIR/res/values/strings.xml|grep '<string'|\
sed 's/.*name="\?//; s/"\?>/ /;s/<\/string>//'\
> getstring-str
join -1 3 -2 1 --nocheck-order getstring-pub getstring-str|awk '{t=$3;$3=$1;$1=t;print}'|\
awk '{id=$2;$1="";$2="";$3="";sub(/^ */,"");printf("s#"id"#\"%s\"#\n", $0)}' > getstring-sed
rm getstring-pub getstring-str
SetStrings
find $1 -name "*.jad" -exec sed -i -f getstring-sed {} \;
Example
Typical example of reverse-engineering and modding APK:
- in APK-Multi-Tool-Linux working dir:
- Drop myapp.apk in place-apk-here-for-modding/
- ./script.sh (and leave it always open in a separate window)
- 9 decompile / 1 select myapp.apk / 9 decompile
- ./getstrings
- Copy apk to dex2jar working dir
- Copy getstring-sed to jad working dir
- in dex2jar working dir:
- ./d2j-dex2jar.sh myapp.apk
- Copy myapp-dex2jar.jar to jad working dir (and/or jd-gui)
- in jad working dir:
- ./unjar myapp-dex2jar.jar
- ./setstrings.sh myapp-dex2jar
- Analyse .jad file and understand what to modify
- in jd-gui working dir:
- As alternative analysis can also be done with jd-gui directly on .jar file
- in APK-Multi-Tool-Linux working dir:
- In working/ find corresponding .smali file and modify it
- (in script.sh windows) 13 compile/sign/install
Dare
Dalvik Retargeting, a tool for converting Android’s .dex format to Java’s .class format
Retargeted .class:
./dare -d output_dir -e myapp.apk
Optimized retargeted .class: (using Soot, slow!)
./dare -o -d output_dir -e myapp.apk
Decompiled optimized retargeted .class: (using Soot, very slow!)
./dare -c -d output_dir -e myapp.apk
APKInspector
The goal of this project is to help analysts and reverse engineers to visualize compiled Android packages and their corresponding DEX code. APKInspector provides both analysis functions and graphic features for the users to gain deep insight into the malicious apps
Still beta and inactive for a year.
GUI around other tools
Androguard
Reverse engineering, Malware analysis of Android applications … and more !
Dexdump
Java .dex file format decompiler
Inactive since 2009
FlowDroid
FlowDroid is a context-, flow-, field-, object-sensitive and lifecycle-aware static taint analysis tool for Android applications
Mobile Sandbox
Provides online static analysis of malware images.
Java
JAD
Java Decompiler
To use on a jar (from dex2jar):
#!/bin/bash
JAD=$(pwd)/jad
ODIR=${1%.jar}
if [ "$ODIR" == "$1" ]; then
echo "Error: expecting a file ending with .jar"
exit 1
fi
7z x -o${ODIR} $1
for d in $(find ${ODIR}/com -type d); do
echo Entering $d
cd $d
for c in *.class; do
$JAD $c
done
cd -
done
./unjar myapp-dex2jar.jar
jadretro
Helps converting Java 1.4, Java 1.5 or later classes so JAD gives better results
JadAlign
Aligns java-files, which are decompiled by jad
java -jar JadHelper-0.0.1.jar myfile.java
No much effect on jad from dex
Jd-gui
JD-GUI is a standalone graphical utility that displays Java source codes of “.class” files
binary-refactor
Helper to manual de-obfuscate obfuscated jars
- rename class/packages in a jar
- match a jarjar-ed & obfuscated jar with a known jar,to find the 'same' classes
- bytecode dump(asm)
- class dependency graph
ELF
man elf
readelf
readelf -a -g -t --dyn-syms -W mybin
elfedit
objdump
objdump -C -g -F -x -T --special-syms mybin objdump -d -l -r -R -S mybin objdump -D -l -r -R -S mybin
nm
nm -a -C -S -s --special-syms mybin
ldd
Shared library dependencies:
ldd -v mybin
PE
Pefile
A Python module to read and work with PE (Portable Executable) files, see usage examples
#!/usr/bin/env python
import sys, pefile
pe = pefile.PE(sys.argv[1])
pe.dump_info()
open('out.txt', 'w').write(pe.dump_info())
Can run under Linux
PEiD
Can run with Wine
PETools
Can run with Wine
Resource Hacker
Can run with Wine
Dependency Walker
Can run with Wine
PEview
Can run with Wine
DLL Export Viewer
Can run with Wine
Under Wine, require absolute path to DLL so: click on gears, "load functions from the following DLL file", Browse
PEBrowse Pro
Can run with Wine
Explorer Suite
- CFF Explorer: Allows also to modify a PE
- Signature Explorer
- PE Detective
- Task Explorer (32 & 64)
PE Insider
Static protections
Packers
- http://www.openrce.org/reference_library/packer_database
- http://www.reverse-engineering.info/documents/33.html
- UPX
upx -d myfile
- Crinkler: some insane PE packing tool coming from the demoscene world.
Dynamic Analysis Tools
IDA Pro
IDA Pro has some debugging capabilities too.
Local debugging: win32, windbg
Remote debugging:
gdbserver --multi <client_ip>:<port> # default IDA port: 23946
Then on IDA: select Remote GDB debugger, paths should be paths on the gdbserver host.
Tuning:
- Debugger / options / Stop on process entry point
- Compatible with lib preloading, cf below
- from 6.4, can make use of Intel PIN tools for diff debugging, see tutorial (pdf)
Intel PIN tools
- Official page
- Windows, Linux, Mac OS X, Android
- x86-32, x86-64 (only Intel platforms obviously)
- binary instrumentation
The best way to think about Pin is as a "just in time" (JIT) compiler. The input to this compiler is not bytecode, however, but a regular executable. Pin intercepts the execution of the first instruction of the executable and generates ("compiles") new code for the straight line code sequence starting at this instruction. It then transfers control to the generated sequence. The generated code sequence is almost identical to the original one, but Pin ensures that it regains control when a branch exits the sequence. After regaining control, Pin generates more code for the branch target and continues execution. Pin makes this efficient by keeping all of the generated code in memory so it can be reused and directly branching from one sequence to another. In JIT mode, the only code ever executed is the generated code. The original code is only used for reference. When generating code, Pin gives the user an opportunity to inject their own code (instrumentation).
Vdb/Vtrace / Vivisect
- debugger, static analysis
- Windows, Linux, Android
- Intel, ARM
vtrace is a cross-platform process debugging API implemented in python, and vdb is a debugger which uses it
vivisect is a Python based static analysis and emulation framework
Android
Binary Instrumentation Framework for Android
Slides here
DroidScope
DECAF(short for Dynamic Executable Code Analysis Framework) is a binary analysis platform based on QEMU.
This is also the home of the DroidScope dynamic Android malware analysis platform. DroidScope is now an extension to DECAF
Slides here and article here
DroidBox
Android Application Sandbox
TaintDroid
Realtime Privacy Monitoring on Smartphones
Soot
Java, Dalvik (see here and here)
GameCIH
GameGuardian
Drozer
Java
Javasnoop
A tool that lets you intercept methods, alter data and otherwise hack Java applications running on your computer.
ELF
ltrace/strace
Tracing library calls and system calls.
Getting a summary:
ltrace -f -S mybin 2>&1|grep '(.*)'|sed 's/(.*//'|sort|uniq -c
Getting more:
ltrace -f -i -S -n 4 -s 1024 mybin
Lib preloading
#define _GNU_SOURCE
#include <dlfcn.h>
#include <sys/types.h>
#include <unistd.h>
#include <errno.h>
#include <stdio.h>
#include <time.h>
// Kill nanosleep()
int nanosleep(const struct timespec *req, struct timespec *rem){
printf("\n==== In our own nanosleep(), I dunnah want sleep\n");
return 0;
}
// Kill usleep()
int usleep(useconds_t usec){
printf("\n==== In our own usleep(), I dunnah want sleep\n");
return 0;
}
// Fix time()
time_t time(time_t *t){
printf("\n==== In our own time(), will return 1380120175\n");
return 1380120175;
}
// Fix srand()
void srand(unsigned int seed){
printf("\n==== In our own srand(), will do srand(0)\n");
void (*original_srand)(unsigned int seed);
original_srand = dlsym(RTLD_NEXT, "srand");
unsigned int myseed = 0;
return (*original_srand)(myseed);
}
#if 0
// Kill rand()
int rand(void){
printf("\n==== In our own rand(), will return 0\n");
return 0;
}
#else
// Intercept rand()
int rand(void){
int (*original_rand)(void);
original_rand = dlsym(RTLD_NEXT, "rand");
int r = (*original_rand)();
printf("\n==== In our own rand(), will return %04X\n", r);
return r;
}
#endif
gcc -fPIC -shared -Wl,-soname,patch -o patch.so patch.c -ldl export LD_PRELOAD=patch.so export LD_LIBRARY_PATH=.:$LD_LIBRARY_PATH
injectso
- x86-32, x86-64, ARM (since v0.52)
scanmem
scanmem is a simple interactive debugging utility for linux, used to locate the address of a variable in an executing process. This can be used for the analysis or modification of a hostile process on a compromised machine, reverse engineering, or as a "pokefinder" to cheat at video games.
- Linux/Android
- with a GUI since v0.13: GameConqueror
PE
Process Monitor
Process Explorer
RegShot
Computes diff between two registry snapshots
HeapMemView
OllyDBG
PE32 only dynamic disassembler and debugger: http://ollydbg.de/.
Support sofwtare and hardware breakpoint, binary patching and repacking, symbol analysis, advanced instruction pattern search, trace with conditional breaking, etc.
There is also a patched version with advanced python scripting ability called Immunity Debugger: http://www.immunityinc.com/products-immdbg.shtml
WinAppDbg
The WinAppDbg python module allows developers to quickly code instrumentation scripts in Python under a Windows environment.
Tracer.py
Based on WinAppDbg, finds interesting bits in trace by dichotomy signal/noise
- run first time and try everything but not the interesting stuff -> use noise option
- then run again and try interesting stuff -> use signal option
WTFDLL.py
Find libraries loaded at runtime and the functions called
Cuckoo Sandboxing
Currently only supporting Windows binaries.
Cuckoo Sandbox is a malware analysis system. You can throw any suspicious file at it and in a matter of seconds Cuckoo will provide you back some detailed results outlining what such file did when executed inside an isolated environment.
Cuckoo generates a handful of different raw data which include:
- Native functions and Windows API calls traces
- Copies of files created and deleted from the filesystem
- Dump of the memory of the selected process
- Full memory dump of the analysis machine
- Screenshots of the desktop during the execution of the malware analysis
- Network dump generated by the machine used for the analysis