Receber por email
← Voltar para o catálogo

reverse-engineering

Binary analysis, disassembly, decompilation, firmware RE, protocol reverse engineering, anti-reversing bypass, malware unpacking

288estrelas
Atualizado há 5 meses

Ver no GitHub ↗ Licença: MIT

Como adicionar

/plugin marketplace add hypnguyen1209/offensive-claude

O comando exato pode variar conforme o repositório. Confira o README no GitHub.

Para o autor da skill

Cole no README do seu repo

Mostra que sua skill está catalogada na Skillteca, gera backlink e tráfego rastreável.

[![Skillteca](https://www.skillteca.com.br/api/badge/reverse-engineering/svg)](https://www.skillteca.com.br/skills/reverse-engineering)

Skills relacionadas

Ver todas de Segurança →

Alerta por categoria

Receba novas skills de Segurança toda segunda

1 email curto, só com as skills novas de Segurança. 4 minutos de leitura, sem spam, cancela com 1 clique.

Você confirma o email no primeiro envio. Sem spam. Cancela com 1 clique.

Ler descrição completa

Reverse Engineering

When to Activate

  • Analyzing compiled binaries for vulnerabilities
  • Understanding proprietary protocols or file formats
  • Malware analysis and unpacking
  • Firmware extraction and analysis
  • Bypassing anti-debugging/anti-tampering protections
  • CTF binary challenges
  • Patch diffing to find 1-day vulnerabilities

Static Analysis

Initial Triage

# File identification
file target_binary
rabin2 -I target_binary  # binary info (arch, bits, endian, protections)

# Strings extraction
strings -n 8 target_binary | grep -iE '(password|key|secret|flag|http|/bin)'
rabin2 -z target_binary   # strings with addresses
rabin2 -zz target_binary  # all strings including wide

# Imports/Exports
rabin2 -i target_binary   # imports
rabin2 -E target_binary   # exports
objdump -T target_binary  # dynamic symbols

# Security mitigations
checksec --file=target_binary
# RELRO, Stack Canary, NX, PIE, FORTIFY

Disassembly & Decompilation

# Ghidra headless analysis
analyzeHeadless /tmp/ghidra_project proj -import target_binary \
  -postScript ExportDecompilation.java -scriptPath /scripts/

# radare2 analysis
r2 -A target_binary
> afl          # list functions
> axt @sym.target_func  # xrefs to function
> pdf @main    # disassemble function
> VV @main     # visual graph mode
> afn new_name @addr  # rename function

# IDA (via MCP or IDAPython)
# Decompile function, rename variables, set types
# Cross-references: xrefs_to(addr), xrefs_from(addr)

Pattern Recognition

# Common vulnerability patterns in disassembly:
# - strcpy/sprintf without bounds → buffer overflow
# - malloc(user_controlled_size) → integer overflow
# - free() followed by use → UAF
# - system()/exec() with user data → command injection
# - Custom crypto (XOR loops, fixed keys) → weak encryption

Dynamic Analysis

Debugging

# GDB with pwndbg/GEF
gdb -q ./target
> break *main
> run
> vmmap              # memory layout
> heap              # heap state
> telescope $rsp 20 # stack inspection
> search-pattern "AAAA"  # find pattern in memory

# Conditional breakpoints
> break *0x401234 if $rax == 0x41414141
> commands
>   x/s $rdi
>   continue
> end

# Anti-debug bypass
> catch syscall ptrace
> commands
>   set $rax = 0
>   continue
> end

Frida Instrumentation

// Hook function and modify behavior
Interceptor.attach(Module.findExportByName(null, "strcmp"), {
    onEnter: function(args) {
        console.log("strcmp(" + args[0].readUtf8String() + ", " + args[1].readUtf8String() + ")");
    },
    onLeave: function(retval) {
        retval.replace(0); // force match
    }
});

// Bypass SSL pinning (Android)
Java.perform(function() {
    var TrustManager = Java.use('com.android.org.conscrypt.TrustManagerImpl');
    TrustManager.verifyChain.implementation = function() {
        return arguments[0];
    };
});

// Trace all JNI calls
Java.perform(function() {
    var System = Java.use('java.lang.System');
    System.loadLibrary.implementation = function(lib) {
        console.log("Loading: " + lib);
        this.loadLibrary(lib);
    };
});

Symbolic Execution

import angr, claripy

proj = angr.Project('./target', auto_load_libs=False)
state = proj.factory.entry_state()

# Symbolic input
sym_input = claripy.BVS('input', 8 * 32)
state.memory.store(input_addr, sym_input)

# Explore to find path to target
simgr = proj.factory.simulation_manager(state)
simgr.explore(find=target_addr, avoid=avoid_addrs)

if simgr.found:
    solution = simgr.found[0].solver.eval(sym_input, cast_to=bytes)
    print(f"Input: {solution}")

Firmware Analysis

# Extraction
binwalk -e firmware.bin
# Filesystem extraction
binwalk --dd='.*' firmware.bin
unsquashfs squashfs-root.img

# Identify architecture
file extracted_binary
readelf -h extracted_binary

# Emulation
qemu-system-arm -M versatilepb -kernel zImage -dtb vexpress.dtb -drive file=rootfs.img

# Common targets in firmware:
# - /etc/shadow, /etc/passwd (hardcoded creds)
# - Web server configs (lighttpd, uhttpd)
# - init scripts (startup services)
# - Proprietary binaries (custom protocols)
# - Certificate/key files

Anti-Reversing Bypass

| Technique | Bypass | |-----------|--------| | IsDebuggerPresent | Patch return value, hook API | | ptrace(PTRACE_TRACEME) | LD_PRELOAD hook, patch syscall | | Timing checks (rdtsc) | Patch comparison, single-step with HW breakpoints | | Self-modifying code | Dump after unpacking, trace execution | | VM detection | Patch CPUID, hide VM artifacts | | Obfuscation (OLLVM) | Symbolic execution, pattern matching, devirtualization | | Packed binaries | Run until OEP, dump from memory | | Anti-disassembly | Fix control flow, NOP junk bytes |

Patch Diffing (1-day Research)

# BinDiff / Diaphora workflow:
# 1. Get vulnerable version and patched version
# 2. Generate IDB/BinExport for both
# 3. Diff — focus on changed functions
# 4. Analyze what was fixed → understand the vulnerability
# 5. Write exploit for the pre-patch version

# Key indicators in patches:
# - Added bounds checks → buffer overflow
# - Added NULL checks → null deref / UAF
# - Changed comparison logic → auth bypass
# - Added sanitization → injection
# - Changed allocation size → heap overflow

Advanced: Firmware Analysis (UEFI/BIOS)

UEFI Extraction & Analysis

# Extract UEFI firmware
UEFIExtract firmware.bin  # Extracts all volumes, sections, files
# Or: uefi-firmware-parser -e firmware.bin

# Identify DXE drivers (most attack surface)
# DXE drivers run with full hardware access before OS loads
# Vulnerable DXE driver = persistent pre-OS implant

# Common UEFI vulnerability classes:
# - SMM (System Management Mode) callout → ring -2 code execution
# - DXE driver buffer overflow → persistent implant
# - Secure Boot bypass → load unsigned bootloader
# - Variable overflow (NVRAM) → code execution in PEI/DXE phase

# Analyze with Ghidra:
# Load as PE/TE binary, set base address from volume header
# Look for: EFI_BOOT_SERVICES, EFI_RUNTIME_SERVICES calls
# Focus on: SMI handlers (SW SMI dispatch), variable access

Secure Boot Bypass Research

# Secure Boot chain: UEFI → shim → GRUB → kernel
# Attack points:
# 1. Vulnerable signed bootloader (BlackLotus technique)
#    - Find old signed bootloader with known vulnerability
#    - Not yet revoked in DBX (Secure Boot blacklist)
#    - Use it to load unsigned code

# 2. GRUB vulnerabilities (BootHole, CVE-2020-10713)
#    - Buffer overflow in GRUB config parsing
#    - Craft malicious grub.cfg → code execution before kernel

# 3. Shim vulnerabilities
#    - Bypass signature verification in shim loader
#    - Load arbitrary EFI binary

# Check revocation status:
# Parse DBX (Forbidden Signatures Database) from NVRAM
# Compare against known-vulnerable bootloader hashes

Advanced: De-obfuscation Techniques

Control Flow Flattening (OLLVM)

# OLLVM flattens control flow into a switch-based dispatcher:
# Original: if/else/loops → Flattened: while(1) { switch(state) { ... } }
# 
# De-flattening approach:
# 1. Identify dispatcher (switch variable, state assignments)
# 2. Trace state transitions for each case
# 3. Reconstruct original control flow graph
# 4. Tools: D-810 (IDA plugin), SATURN, deflat.py

# Symbolic execution for de-obfuscation:
import angr

def deobfuscate_cfg(binary, func_addr):
    proj = angr.Project(binary, auto_load_libs=False)
    cfg = proj.analyses.CFGFast()
    func = cfg.functions[func_addr]
    
    # Identify dispatcher block (highest in-degree)
    dispatcher = max(func.blocks, key=lambda b: len(list(func.graph.predecessors(b))))
    
    # For each state value, trace execution to find real successor
    # Build de-obfuscated CFG from state transitions

String Deobfuscation (Automated)

# Common patterns:
# 1. XOR with key: for(i=0;i<len;i++) str[i] ^= key[i%keylen]
# 2. Stack strings: mov [rsp+0], 'H'; mov [rsp
Compartilhar X LinkedIn