Binary Exploitation Methodology

Binary Exploitation Methodology

A structured approach to analyzing and exploiting binary vulnerabilities, particularly stack-based buffer overflows and related techniques.

When to Use This Skill

Use this methodology when:

• Analyzing a binary for vulnerabilities
• Developing exploits for CTF challenges
• Researching buffer overflow vulnerabilities
• Bypassing security protections (ASLR, PIE, NX, canaries)
• Planning exploitation strategies for compiled binaries

Phase 1: Binary Analysis

1.1 Gather ELF Information

Before attempting exploitation, understand the binary's structure:

# Basic binary info
file <binary>
readelf -h <binary>

# Check for PIE
readelf -d <binary> | grep -i type

# Check for stack canaries
readelf -s <binary> | grep -i __stack_chk

# Check for NX (non-executable stack)
readelf -l <binary> | grep GNU_STACK

# Check for ASLR
readelf -d <binary> | grep FLAGS

1.2 Identify Protections

Phase 2: Vulnerability Identification

2.1 Common Vulnerability Types

Stack Overflows

• Overwriting return pointer (EIP/RIP)
• Overwriting EBP/RBP to control ESP/RSP
• May require integer overflow to trigger

Arbitrary Write Vulnerabilities

• Format string bugs (printf family)
• Array indexing issues
• Buffer overflow to Write-What-Where via ROP

2.2 Analysis Tools

Use these tools to identify vulnerabilities:

• GDB with GEF/PEDA for debugging
• pwntools for exploit development
• radare2/ghidra for reverse engineering
• checksec for protection analysis

Phase 3: Exploitation Goals

Goal 1: Call an Existing Function (ret2win)

Scenario: A win() function exists that grants access

Without PIE and Canary:

[buffer][padding][return_address -> win_function]

With PIE: Need to leak an address first to calculate base

With Canary: Need to leak or bypass the canary value

Setting Function Parameters:

• Use ROP chain if enough gadgets exist
• Use SROP (sigreturn-oriented programming) to control registers
• Use ret2csu/ret2vdso gadgets for register control

Goal 2: Remote Code Execution (RCE)

Via Shellcode

Prerequisites: NX disabled or bypassed

Without ASLR and NX:

[shellcode][padding][return_address -> shellcode_location]

With ASLR: Use ret2esp/ret2reg to jump to stack

With NX: Use ROP to call mprotect() and make memory executable

Via Syscalls (ret2syscall)

Use case: Call execve() to run arbitrary commands

Requirements:

• Find gadgets to set syscall parameters
• Bypass ASLR/PIE to use ROP gadgets
• Consider SROP for ret2execve

Via libc (ret2libc)

Use case: Call system("/bin/sh") from libc

Scenarios:

ret2libc payload structure:

[padding][return_to_system][fake_return][pointer_to_/bin/sh]

Goal 3: Stack Pivoting

Use case: Off-by-one overflows or alternate EIP control

Technique: Control ESP/RSP through EBP/RBP chain

Phase 4: Eternal Loops

When a single exploitation attempt isn't enough, create loops:

4.1 ROP Chain to Main

Write the address of main() or the vulnerable function in the ROP chain to re-execute the vulnerability.

4.2 GOT Entry Modification

Overwrite exit GOT entry to point back to the vulnerable function.

4.3 .fini_array Abuse

Store two functions in .fini_array:

1. Function to call the vulnerability again
2. Function to call __libc_csu_fini (which re-executes .fini_array)

Phase 5: Protection Bypasses

5.1 Bypassing ASLR

Methods:

• Leak a libc/binary address and calculate base
• Use information disclosure vulnerabilities
• Brute force (only for 32-bit or small address space)

5.2 Bypassing PIE

Methods:

• Leak any function address to calculate base
• Use GOT entries (not affected by PIE)

5.3 Bypassing NX

Methods:

• ROP chain to mprotect() to make memory executable
• Use existing executable memory (code reuse)
• Find gadgets that don't require shellcode

5.4 Bypassing Canary

Methods:

• Leak canary via format string or info disclosure
• Overwrite canary before it's checked
• Use vulnerabilities that don't trigger canary check

Phase 6: Exploit Development Workflow

Step 1: Fuzz and Identify

from pwn import *

# Find offset to return address
offset = cyclic_find(pwn.cyclic(1000))

# Test for vulnerabilities
process = process('./binary')
process.sendline(cyclic(1000))
process.wait()

Step 2: Leak Addresses (if needed)

# Leak libc address via GOT
libc_addr = u64(process.read(8).ljust(8, b'\x00'))
libc_base = libc_addr - libc.symbols['function_name']

Step 3: Build Payload

payload = flat([
    cyclic(offset),           # Fill to return address
    libc.symbols['system'],   # Call system
    0,                        # Fake return
    next(libc.search(b'/bin/sh'))  # /bin/sh address
])

Step 4: Test and Refine

• Test locally first
• Handle timing issues
• Account for different environments

Common Pitfalls

1. Endianness: Remember x86 is little-endian
2. Bad characters: Avoid null bytes, newlines in shellcode
3. Alignment: Some gadgets require stack alignment
4. Timing: Race conditions in multi-process scenarios
5. Environment differences: Test on target architecture

Quick Reference

ROP Chain Structure

[padding][ret][ret][ret][...][final_ret]

ret2libc Structure

[padding][system_addr][fake_ret][/bin/sh_addr]

Shellcode Structure

[nop_sled][shellcode][padding][return_to_shellcode]

Next Steps

After identifying the vulnerability type and protections:

1. Choose the appropriate exploitation goal
2. Select the technique based on protection status
3. Develop and test the exploit
4. Refine for reliability

Remember: Each binary is unique. Adapt this methodology to the specific protections and constraints you encounter.