#!/usr/bin/env python3
"""Scan a sample for destructive wiper primitives and classify wiper vs ransomware.

Detects raw-disk/MBR overwrite primitives, backup/recovery destruction commands, and weighs them
against ransomware (encryption/extortion) indicators to classify intent. Reads bytes statically;
executes nothing.
"""
from __future__ import annotations

import argparse
import json
import re
import sys
from pathlib import Path

ASCII_RE = re.compile(rb"[\x20-\x7e]{4,}")
UTF16_RE = re.compile(rb"(?:[\x20-\x7e]\x00){4,}")

RAW_DISK = ["\\\\.\\physicaldrive", "\\\\.\\c:", "deviceiocontrol", "ioctl_disk",
            "createfilemapping", "setfilepointer", "\\device\\harddisk"]
BOOT_TAMPER = ["this program cannot be run", "mbr", "\\\\.\\physicaldrive0", "bootmgr is missing"]
RECOVERY_DENIAL = ["vssadmin delete shadows", "vssadmin.exe delete", "wbadmin delete",
                   "bcdedit /set", "recoveryenabled no", "wevtutil cl", "delete catalog",
                   "shadowcopy delete"]
MASS_DELETE = ["deletefile", "removedirectory", "shfileoperation", "writefile", "f0 fill",
               "overwrite"]
RANSOM_INDICATORS = ["cryptencrypt", "cryptdecrypt", "bcryptencrypt", "readme.txt", "decrypt",
                     ".onion", "bitcoin", "ransom", "your files have been encrypted",
                     "aes", "rsa public key"]


def _strings(data: bytes) -> str:
    parts = [m.decode("ascii", "ignore") for m in ASCII_RE.findall(data)]
    parts += [m.decode("utf-16-le", "ignore") for m in UTF16_RE.findall(data)]
    return "\n".join(parts).lower()


def scan(path: str) -> dict:
    blob = _strings(Path(path).read_bytes())
    raw = sorted({r for r in RAW_DISK if r in blob})
    boot = sorted({b for b in BOOT_TAMPER if b in blob})
    recovery = sorted({r for r in RECOVERY_DENIAL if r in blob})
    deletion = sorted({d for d in MASS_DELETE if d in blob})
    ransom = sorted({r for r in RANSOM_INDICATORS if r in blob})
    destructive_score = len(raw) * 2 + len(boot) * 2 + len(recovery) + len(deletion)
    ransom_score = len(ransom)
    if destructive_score >= 3 and ransom_score <= 1:
        classification = "likely_wiper"
    elif ransom_score >= 3:
        classification = "likely_ransomware"
    elif destructive_score and ransom_score:
        classification = "destructive_with_ransom_facade"
    else:
        classification = "inconclusive"
    return {
        "raw_disk_primitives": raw,
        "boot_record_tampering": boot,
        "recovery_denial": recovery,
        "mass_deletion": deletion,
        "ransomware_indicators": ransom,
        "destructive_score": destructive_score,
        "ransomware_score": ransom_score,
        "classification": classification,
    }


def main(argv: list[str] | None = None) -> int:
    parser = argparse.ArgumentParser(description="Analyze wiper malware")
    sub = parser.add_subparsers(dest="cmd", required=True)
    p = sub.add_parser("scan")
    p.add_argument("file")
    args = parser.parse_args(argv)
    print(json.dumps(scan(args.file), indent=2))
    return 0


if __name__ == "__main__":
    sys.exit(main())