Reverse Engineering Rust Malware

Reverse Engineering Rust Malware

Overview

Rust has become increasingly popular for malware development due to its cross-compilation, memory safety guarantees, and the complexity it introduces for reverse engineers. Rust binaries contain the entire standard library statically linked, producing large binaries with extensive boilerplate code. Key challenges include non-null-terminated strings (Rust uses fat pointers with pointer+length), monomorphization generating duplicated generic code, complex error handling (Result/Option unwrap chains), and unfamiliar calling conventions. Decompiling Rust to C produces unhelpful output compared to C/C++ binaries. Tools like Ghidra scripts for crate extraction, and training focused on Rust-specific patterns (2024-2025) help address these challenges. Notable Rust malware includes BlackCat/ALPHV ransomware, Hive ransomware variants, and Buer Loader.

When to Use

• When performing authorized security testing that involves reverse engineering rust malware
• When analyzing malware samples or attack artifacts in a controlled environment
• When conducting red team exercises or penetration testing engagements
• When building detection capabilities based on offensive technique understanding

Detection Gaps & Validation

• Stripped binaries erase the easy wins. Release-mode Rust with strip and panic=abort removes panicked at/core::panicking strings and .cargo/registry paths, so the identify-and-extract-crates step returns almost nothing. Absence of these markers is not "not Rust" — fall back to detecting the std runtime (lang_start), the allocator, and monomorphized generic patterns.
• Non-null-terminated fat-pointer strings break normal extraction. Rust stores &str as pointer+length with no terminator, so adjacent literals run together in .rodata and a C-string strings dump produces garbled concatenations. Recover boundaries by reading the length from the slice's fat pointer at the use site, not by scanning for \x00.
• Decompilation is genuinely poor — verify at the assembly level. Monomorphization duplicates generic code, Result/Option unwrap chains and bounds-check panics bloat every function, and Ghidra's C output is misleading. Confirm logic in disassembly and use the crate list (reqwest/tokio/aes/winreg) to anchor where networking/crypto/persistence actually live.
• Crate inference ≠ proof of behavior. Seeing aes or reqwest linked tells you a capability may exist; static linking pulls in unused code paths. Confirm the function is reached before reporting it.
• Confirm dynamically: for stripped/obfuscated samples, capture decrypted strings and C2 at runtime (x64dbg/IDA debugger, breakpoint after the decrypt/format call) and a network capture — static-only analysis of Rust malware is the top false-negative source.
• Benign-lookalike FPs: large size, statically-linked std, and crypto crates are normal for all Rust binaries (including legitimate tooling); never flag on binary size or crate presence alone.

Prerequisites

• IDA Pro 8.0+ or Ghidra 11.0+
• Rust toolchain for reference compilation
• Python 3.9+ for helper scripts
• Understanding of Rust memory model (ownership, borrowing)
• Familiarity with Rust string types (String, &str, CString)

Workflow

Step 1: Identify and Parse Rust Binary Metadata

#!/usr/bin/env python3
"""Analyze Rust malware binary metadata and extract crate dependencies."""
import re
import sys
import json


def identify_rust_binary(data):
    """Check if binary is Rust-compiled and extract version info."""
    indicators = {
        "rust_panic_strings": bool(re.search(rb'panicked at', data)),
        "rust_unwrap": bool(re.search(rb'called.*unwrap.*on.*None', data)),
        "core_panic": bool(re.search(rb'core::panicking', data)),
        "std_rt": bool(re.search(rb'std::rt::lang_start', data)),
        "cargo_path": bool(re.search(rb'\.cargo[/\\]registry', data)),
        "rustc_version": None,
    }

    version = re.search(rb'rustc\s+(\d+\.\d+\.\d+)', data)
    if version:
        indicators["rustc_version"] = version.group(1).decode()

    is_rust = sum(1 for v in indicators.values() if v) >= 2
    return is_rust, indicators


def extract_crates(data):
    """Extract Rust crate (dependency) names from binary strings."""
    crate_pattern = re.compile(
        rb'(?:crates\.io-[a-f0-9]+/|\.cargo/registry/src/[^/]+/)'
        rb'([\w-]+)-(\d+\.\d+\.\d+)'
    )
    crates = {}
    for match in crate_pattern.finditer(data):
        name = match.group(1).decode()
        version = match.group(2).decode()
        crates[name] = version

    # Also check for common malware-relevant crates
    suspicious_crates = {
        "reqwest": "HTTP client",
        "hyper": "HTTP library",
        "tokio": "Async runtime",
        "aes": "AES encryption",
        "chacha20": "ChaCha20 encryption",
        "rsa": "RSA encryption",
        "ring": "Crypto library",
        "base64": "Base64 encoding",
        "winapi": "Windows API bindings",
        "winreg": "Registry access",
        "sysinfo": "System information",
        "screenshots": "Screen capture",
        "clipboard": "Clipboard access",
        "keylogger": "Key logging",
    }

    capabilities = []
    for crate_name, description in suspicious_crates.items():
        if crate_name in crates:
            capabilities.append({
                "crate": crate_name,
                "version": crates[crate_name],
                "capability": description,
            })

    return crates, capabilities


def extract_rust_strings(data):
    """Extract strings handling Rust's non-null-terminated format."""
    # Rust strings are stored as pointer+length, but string literals
    # are often in .rodata as contiguous sequences
    strings = []
    ascii_pattern = re.compile(rb'[\x20-\x7e]{8,500}')
    for match in ascii_pattern.finditer(data):
        s = match.group().decode('ascii')
        # Filter for malware-relevant strings
        keywords = ['http', 'socket', 'encrypt', 'decrypt', 'shell',
                    'exec', 'cmd', 'upload', 'download', 'persist',
                    'registry', 'mutex', 'pipe', 'inject']
        if any(kw in s.lower() for kw in keywords):
            strings.append(s)

    return strings


if __name__ == "__main__":
    if len(sys.argv) < 2:
        print(f"Usage: {sys.argv[0]} <rust_binary>")
        sys.exit(1)

    with open(sys.argv[1], 'rb') as f:
        data = f.read()

    is_rust, indicators = identify_rust_binary(data)
    print(f"[{'+'if is_rust else '-'}] Rust binary: {is_rust}")
    print(json.dumps(indicators, indent=2, default=str))

    crates, capabilities = extract_crates(data)
    print(f"\n[+] Crates ({len(crates)}):")
    for name, ver in sorted(crates.items()):
        print(f"  {name} v{ver}")

    if capabilities:
        print(f"\n[!] Suspicious capabilities:")
        for cap in capabilities:
            print(f"  {cap['crate']} -> {cap['capability']}")

    strings = extract_rust_strings(data)
    if strings:
        print(f"\n[+] Suspicious strings ({len(strings)}):")
        for s in strings[:20]:
            print(f"  {s}")

Validation Criteria

• Binary correctly identified as Rust-compiled with version info
• Crate dependencies extracted revealing malware capabilities
• Rust-specific string extraction handles fat pointer format
• Main entry point and core logic functions identified
• Encryption, networking, and persistence code located

References

• Binary Defense - Extracting Secrets from Rust Malware
• Ghidra Extension for Rust Analysis
• Fuzzing Labs - Reversing Modern Binaries
• Bishop Fox - Rust for Malware Development