Active Directory — SPN Unicode Collision Detection Scanner (CVE-2026-25177)

  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500

#!/usr/bin/env python3
"""
CVE-2026-25177 Detector - Safe for Production Use
===================================================
Read-only scanner that detects signs of SPN Unicode collision attacks
against Active Directory Domain Controllers.

This script performs ONLY read operations (LDAP searches). It does NOT
modify any AD objects, write any attributes, or change any configuration.

Detects:
  1. SPNs containing Unicode zero-width or homoglyph characters
  2. Duplicate SPNs across multiple accounts (case-insensitive)
  3. SPNs that visually match but differ at the byte level
  4. Recent SPN modifications (via whenChanged attribute)

Requirements:
  pip install ldap3 pycryptodome

Usage:
  python detect_spn_abuse.py                                    # auto-detect everything, prompt for password
  python detect_spn_abuse.py -dc 10.0.0.1 -d corp.local        # prompt for user and password
  python detect_spn_abuse.py -dc 10.0.0.1 -d corp.local -u svc_scanner --csv report.csv --days 7
"""

import argparse
import csv
import getpass
import os
import socket
import ssl
import subprocess
import sys
import unicodedata
from collections import defaultdict
from datetime import datetime, timedelta

from ldap3 import Server, Connection, ALL, NTLM, SIMPLE, Tls, SUBTREE


# Characters that should never appear in a legitimate SPN
SUSPICIOUS_CODEPOINTS = {
    0x200B: "Zero-Width Space",
    0x200C: "Zero-Width Non-Joiner",
    0x200D: "Zero-Width Joiner",
    0x200E: "Left-to-Right Mark",
    0x200F: "Right-to-Left Mark",
    0x2060: "Word Joiner",
    0x2061: "Function Application",
    0x2062: "Invisible Times",
    0x2063: "Invisible Separator",
    0x2064: "Invisible Plus",
    0xFEFF: "Byte Order Mark / ZWNBSP",
    0xFFFE: "Reversed BOM",
    0x00AD: "Soft Hyphen",
    0x034F: "Combining Grapheme Joiner",
    0x180E: "Mongolian Vowel Separator",
    0x061C: "Arabic Letter Mark",
    0x115F: "Hangul Choseong Filler",
    0x1160: "Hangul Jungseong Filler",
    0x17B4: "Khmer Vowel Inherent Aq",
    0x17B5: "Khmer Vowel Inherent Aa",
}

# Homoglyph pairs: (unicode char, ascii char it mimics)
HOMOGLYPH_MAP = {
    0x2044: ("/", "Fraction Slash"),
    0xFF0F: ("/", "Fullwidth Solidus"),
    0x2215: ("/", "Division Slash"),
    0xFF28: ("H", "Fullwidth H"),
    0xFF34: ("T", "Fullwidth T"),
    0xFF30: ("P", "Fullwidth P"),
    0x0391: ("A", "Greek Alpha"),
    0x0392: ("B", "Greek Beta"),
    0x0395: ("E", "Greek Epsilon"),
    0x0397: ("H", "Greek Eta"),
    0x0399: ("I", "Greek Iota"),
    0x039A: ("K", "Greek Kappa"),
    0x039C: ("M", "Greek Mu"),
    0x039D: ("N", "Greek Nu"),
    0x039F: ("O", "Greek Omicron"),
    0x03A1: ("P", "Greek Rho"),
    0x03A4: ("T", "Greek Tau"),
    0x03A5: ("Y", "Greek Upsilon"),
    0x0410: ("A", "Cyrillic A"),
    0x0412: ("B", "Cyrillic Ve"),
    0x0415: ("E", "Cyrillic Ie"),
    0x041D: ("H", "Cyrillic En"),
    0x041E: ("O", "Cyrillic O"),
    0x0420: ("P", "Cyrillic Er"),
    0x0421: ("C", "Cyrillic Es"),
    0x0422: ("T", "Cyrillic Te"),
}


def auto_discover():
    """Auto-discover DC and domain from the current machine.
    Works on domain-joined Windows machines."""
    dc = None
    domain = None

    # Method 1: Environment variables
    domain = os.environ.get("USERDNSDOMAIN")
    logon_server = os.environ.get("LOGONSERVER", "").strip("\\")

    if domain and logon_server:
        try:
            dc = socket.gethostbyname(logon_server)
            print(f"[+] Auto-discovered from environment:")
            print(f"    Domain:  {domain}")
            print(f"    DC:      {logon_server} ({dc})")
            return dc, domain
        except socket.gaierror:
            pass

    # Method 2: nltest
    try:
        result = subprocess.run(
            ["nltest", "/dsgetdc:"],
            capture_output=True, text=True, timeout=5
        )
        if result.returncode == 0:
            for line in result.stdout.splitlines():
                line = line.strip()
                if line.startswith("DC:"):
                    dc_name = line.split("DC:")[1].strip().strip("\\")
                    try:
                        dc = socket.gethostbyname(dc_name)
                    except socket.gaierror:
                        dc = dc_name
                elif "Dns Dom Name:" in line or "DNS Dom Name:" in line:
                    domain = line.split(":")[1].strip()
                elif line.startswith("Dom Name:") and not domain:
                    domain = line.split("Dom Name:")[1].strip()

            if dc and domain:
                print(f"[+] Auto-discovered via nltest:")
                print(f"    Domain:  {domain}")
                print(f"    DC:      {dc}")
                return dc, domain
    except (FileNotFoundError, subprocess.TimeoutExpired):
        pass

    # Method 3: DNS SRV lookup
    if domain:
        try:
            result = subprocess.run(
                ["nslookup", "-type=SRV", f"_ldap._tcp.dc._msdcs.{domain}"],
                capture_output=True, text=True, timeout=5
            )
            for line in result.stdout.splitlines():
                if "svr hostname" in line.lower():
                    dc_name = line.split("=")[-1].strip().rstrip(".")
                    try:
                        dc = socket.gethostbyname(dc_name)
                    except socket.gaierror:
                        dc = dc_name
                    print(f"[+] Auto-discovered via DNS SRV:")
                    print(f"    Domain:  {domain}")
                    print(f"    DC:      {dc_name} ({dc})")
                    return dc, domain
        except (FileNotFoundError, subprocess.TimeoutExpired):
            pass

    return dc, domain


def connect_readonly(dc_host, domain, username, password):
    """Establish a read-only LDAP connection."""
    tls_config = Tls(validate=ssl.CERT_NONE)
    methods = [
        ("NTLM:389", 389, False, NTLM, f"{domain}\\{username}"),
        ("NTLM:636", 636, True, NTLM, f"{domain}\\{username}"),
        ("SIMPLE:636", 636, True, SIMPLE, f"{username}@{domain}"),
        ("SIMPLE:389", 389, False, SIMPLE, f"{username}@{domain}"),
    ]
    for name, port, use_ssl, auth, user in methods:
        try:
            server = Server(dc_host, port=port, use_ssl=use_ssl,
                            tls=tls_config if use_ssl else None,
                            get_info=ALL)
            conn = Connection(server, user=user, password=password,
                              authentication=auth, auto_bind=True,
                              read_only=True)
            print(f"[+] Connected via {name} (read-only)")
            return conn
        except Exception:
            continue
    print("[-] All authentication methods failed")
    sys.exit(1)


def normalize_spn(spn):
    """Normalize an SPN for comparison: strip non-ASCII, lowercase."""
    cleaned = ""
    for ch in spn:
        cp = ord(ch)
        if cp in SUSPICIOUS_CODEPOINTS:
            continue
        if cp in HOMOGLYPH_MAP:
            cleaned += HOMOGLYPH_MAP[cp][0]
            continue
        if cp > 127:
            # Try NFKD normalization (decomposes fullwidth etc.)
            normalized = unicodedata.normalize("NFKD", ch)
            if all(ord(c) < 128 for c in normalized):
                cleaned += normalized
                continue
        cleaned += ch
    return cleaned.lower()


def check_unicode_chars(spn):
    """Check an SPN for suspicious Unicode characters. Returns list of findings."""
    findings = []

    for i, ch in enumerate(spn):
        cp = ord(ch)

        # Check zero-width / invisible characters
        if cp in SUSPICIOUS_CODEPOINTS:
            findings.append({
                "position": i,
                "char": ch,
                "codepoint": f"U+{cp:04X}",
                "type": "invisible",
                "name": SUSPICIOUS_CODEPOINTS[cp],
            })

        # Check homoglyphs
        elif cp in HOMOGLYPH_MAP:
            ascii_equiv, name = HOMOGLYPH_MAP[cp]
            findings.append({
                "position": i,
                "char": ch,
                "codepoint": f"U+{cp:04X}",
                "type": "homoglyph",
                "name": f"{name} (looks like '{ascii_equiv}')",
            })

        # Check any other non-ASCII (SPNs should be ASCII-only)
        elif cp > 127:
            cat = unicodedata.category(ch)
            char_name = unicodedata.name(ch, "UNKNOWN")
            findings.append({
                "position": i,
                "char": ch,
                "codepoint": f"U+{cp:04X}",
                "type": "non-ascii",
                "name": f"{char_name} (category: {cat})",
            })

    return findings


def scan_spns(conn, search_base, days_back=None):
    """Scan all SPNs in the directory. Returns (all_spns, alerts)."""
    print("[*] Scanning all accounts with SPNs...")

    attrs = ["sAMAccountName", "servicePrincipalName",
             "distinguishedName", "whenChanged", "objectClass"]

    conn.search(search_base, "(servicePrincipalName=*)",
                search_scope=SUBTREE, attributes=attrs,
                paged_size=500, paged_cookie=None)

    # Handle LDAP paging to get ALL results (default AD limit is 1000)
    all_entries = list(conn.entries)
    page_num = 1
    print(f"    Page {page_num}: retrieved {len(all_entries)} accounts...", flush=True)
    while conn.result.get("controls", {}).get("1.2.840.113556.1.4.319", {}).get("value", {}).get("cookie"):
        cookie = conn.result["controls"]["1.2.840.113556.1.4.319"]["value"]["cookie"]
        conn.search(search_base, "(servicePrincipalName=*)",
                    search_scope=SUBTREE, attributes=attrs,
                    paged_size=500, paged_cookie=cookie)
        all_entries.extend(conn.entries)
        page_num += 1
        print(f"    Page {page_num}: retrieved {len(all_entries)} accounts so far...", flush=True)

    print(f"    Done. {len(all_entries)} total accounts to analyze.")

    all_spns = []       # list of (account, dn, spn, when_changed)
    alerts = []         # list of alert dicts
    spn_index = defaultdict(list)  # normalized_spn -> [(account, spn, dn)]

    cutoff = None
    if days_back:
        from datetime import timezone
        cutoff = datetime.now(timezone.utc) - timedelta(days=days_back)

    total_accounts = len(all_entries)
    total_spns = 0

    for entry in all_entries:
        account = str(entry.sAMAccountName)
        dn = str(entry.distinguishedName)
        spns = list(entry.servicePrincipalName) if entry.servicePrincipalName else []
        when_changed = entry.whenChanged.value if hasattr(entry, "whenChanged") and entry.whenChanged.value else None
        obj_classes = list(entry.objectClass) if entry.objectClass else []

        for spn in spns:
            total_spns += 1
            all_spns.append((account, dn, spn, when_changed))

            # Index by normalized form for duplicate detection
            norm = normalize_spn(spn)
            spn_index[norm].append((account, spn, dn))

            # Check 1: Unicode characters in SPN
            unicode_findings = check_unicode_chars(spn)
            if unicode_findings:
                alerts.append({
                    "check": "unicode_chars",
                    "severity": "CRITICAL",
                    "account": account,
                    "dn": dn,
                    "spn": spn,
                    "spn_hex": spn.encode("utf-8").hex(),
                    "details": unicode_findings,
                    "when_changed": when_changed,
                    "message": f"SPN contains {len(unicode_findings)} suspicious Unicode character(s)",
                })

            # Check 2: Recently modified SPNs (if days_back specified)
            if cutoff and when_changed:
                # Make comparison timezone-aware
                from datetime import timezone
                if when_changed.tzinfo is None:
                    when_changed = when_changed.replace(tzinfo=timezone.utc)
                if when_changed > cutoff:
                    # Only alert on user accounts, not computer accounts
                    is_computer = "computer" in [c.lower() for c in obj_classes]
                    if not is_computer:
                        alerts.append({
                            "check": "recent_change",
                            "severity": "INFO",
                            "account": account,
                            "dn": dn,
                            "spn": spn,
                            "when_changed": when_changed,
                            "message": f"SPN on user account modified within last {days_back} days",
                        })

    # Check 3: Duplicate SPNs (different accounts, same normalized SPN)
    for norm_spn, entries in spn_index.items():
        if len(entries) > 1:
            accounts_involved = [(a, s, d) for a, s, d in entries]
            # Check if the raw SPNs differ (Unicode collision indicator)
            raw_spns = set(s for _, s, _ in entries)
            if len(raw_spns) > 1:
                severity = "CRITICAL"
                msg = "Byte-level different SPNs resolve to same value (Unicode collision!)"
            else:
                severity = "HIGH"
                msg = "Exact duplicate SPN on multiple accounts"

            alerts.append({
                "check": "duplicate_spn",
                "severity": severity,
                "normalized_spn": norm_spn,
                "accounts": accounts_involved,
                "message": msg,
            })

    print(f"[+] Scanned {total_accounts} accounts, {total_spns} SPNs")
    return all_spns, alerts


def print_report(alerts):
    """Print a formatted report of findings."""
    if not alerts:
        print("\n[+] No suspicious SPN activity detected.")
        return

    # Sort by severity
    severity_order = {"CRITICAL": 0, "HIGH": 1, "MEDIUM": 2, "INFO": 3}
    alerts.sort(key=lambda a: severity_order.get(a["severity"], 99))

    critical = sum(1 for a in alerts if a["severity"] == "CRITICAL")
    high = sum(1 for a in alerts if a["severity"] == "HIGH")
    info = sum(1 for a in alerts if a["severity"] == "INFO")

    print(f"\n{'=' * 70}")
    print(f"  SCAN RESULTS: {len(alerts)} finding(s)")
    print(f"  CRITICAL: {critical}  |  HIGH: {high}  |  INFO: {info}")
    print(f"{'=' * 70}")

    for i, alert in enumerate(alerts):
        sev = alert["severity"]
        check = alert["check"]

        print(f"\n  [{sev}] Finding {i + 1}: {alert['message']}")
        print(f"  {'─' * 60}")

        if check == "unicode_chars":
            print(f"  Account:  {alert['account']}")
            print(f"  DN:       {alert['dn']}")
            print(f"  SPN:      {alert['spn']}")
            print(f"  SPN Hex:  {alert['spn_hex']}")
            if alert.get("when_changed"):
                print(f"  Modified: {alert['when_changed']}")
            for finding in alert["details"]:
                print(f"    Position {finding['position']}: "
                      f"{finding['codepoint']} - {finding['name']} "
                      f"[{finding['type']}]")

        elif check == "duplicate_spn":
            print(f"  Normalized SPN: {alert['normalized_spn']}")
            print(f"  Accounts with this SPN:")
            for account, raw_spn, dn in alert["accounts"]:
                print(f"    - {account}")
                print(f"      DN:  {dn}")
                print(f"      Raw: {raw_spn}")
                print(f"      Hex: {raw_spn.encode('utf-8').hex()}")

        elif check == "recent_change":
            print(f"  Account:  {alert['account']}")
            print(f"  DN:       {alert['dn']}")
            print(f"  SPN:      {alert['spn']}")
            print(f"  Modified: {alert['when_changed']}")

    print(f"\n{'=' * 70}")

    if critical > 0:
        print("\n  RECOMMENDED ACTIONS:")
        print("  1. Investigate accounts with CRITICAL findings immediately")
        print("  2. Remove any SPNs containing Unicode characters")
        print("  3. Check Event Log for SPN modification events (Event ID 4742)")
        print("  4. Review delegation permissions (who can write SPNs)")
        print("  5. Apply Microsoft patch for CVE-2026-25177")
    print()


def write_csv(alerts, filename):
    """Write findings to a CSV file."""
    with open(filename, "w", newline="", encoding="utf-8") as f:
        writer = csv.writer(f)
        writer.writerow(["Severity", "Check", "Account", "DN", "SPN",
                         "SPN_Hex", "Message", "Modified", "Details"])

        for alert in alerts:
            account = alert.get("account", "")
            dn = alert.get("dn", "")
            spn = alert.get("spn", "")
            spn_hex = alert.get("spn_hex", "")
            when = str(alert.get("when_changed", ""))

            if alert["check"] == "duplicate_spn":
                for a, s, d in alert["accounts"]:
                    writer.writerow([
                        alert["severity"], alert["check"], a, d, s,
                        s.encode("utf-8").hex(), alert["message"], "", ""
                    ])
            else:
                details = ""
                if alert.get("details"):
                    details = "; ".join(
                        f"pos {f['position']}: {f['codepoint']} {f['name']}"
                        for f in alert["details"]
                    )
                writer.writerow([
                    alert["severity"], alert["check"], account, dn, spn,
                    spn_hex, alert["message"], when, details
                ])

    print(f"[+] CSV report written to: {filename}")


def main():
    parser = argparse.ArgumentParser(
        description="CVE-2026-25177 Detector - Production-safe SPN scanner",
        formatter_class=argparse.RawDescriptionHelpFormatter,
        epilog="""
This scanner is READ-ONLY and safe for production use.
It does not modify any Active Directory objects.

Examples:
  # Auto-detect everything (domain-joined machine), prompt for password
  python detect_spn_abuse.py

  # Auto-detect DC and domain, specify user
  python detect_spn_abuse.py -u svc_scanner

  # Fully manual
  python detect_spn_abuse.py -dc dc01.corp.local -d corp.local -u scanner -p 'pass'

  # Scan with CSV output and recent change detection
  python detect_spn_abuse.py --csv report.csv --days 7

  # Use LDAPS
  python detect_spn_abuse.py --use-ssl
        """,
    )

    parser.add_argument("-dc", required=False,
                        help="Domain Controller IP/hostname (auto-detected if omitted)")
    parser.add_argument("-d", "--domain", required=False,
                        help="Domain name (auto-detected if omitted)")
    parser.add_argument("-u", "--username", required=False,
                        help="Username (auto-detected if omitted)")