Date: Fri, 9 Jul 2010 07:25:08 +0400 From: Solar Designer <solar@...nwall.com> To: john-users@...ts.openwall.com Subject: LM and NTLM C/R cracking Hi, Here's an example of how LM and NTLM challenge/response pairs may be processed with John the Ripper with the jumbo patch applied, using sample C/R pairs from this recent blog post: http://carnal0wnage.attackresearch.com/node/427 http://carnal0wnage.blogspot.com/2010/07/revisiting-halflm-stuff.html I formatted the input file as follows: ADMIN:::59DE5D885E583167C3A9A92AC42C0AE52F85252CC731BB25:5ADA49D539BD174E7049805DC1004925E25130C33DBE892A:1122334455667788 ADMIN:::40305B22075D6000D0508D9AD1F7BEB02F85252CC731BB25:337C939E66480243D1833309B8AFE49A81FE4C5E646BF00A:1122334455667788 ADMIN:::DAF3570C10ED2817C3D8A05D69F9EF292F85252CC731BB25:D3FB390BAC5D152F7A394466FBEF686E275D05B99C0A115E:1122334455667788 ADMIN:::76365E2D142B5612980C67D057EB9EFEEE5EF6EB6FF6E04D:727B4E35F947129EA52B9CDEDAE86934BB23EF89F50FC595:1122334455667788 ADMIN:::D737AA8F95CE38359CAB5D8A2519C4B92F85252CC731BB25:0624A3F7D457C54B163C641DBF4B7963548EF1C5D0397CBF:1122334455667788 ADMIN:::0E89A68D07E315C6035E82B757B955882F85252CC731BB25:58F2D720179B4A38A0523E02AEF0D41DACCCD6577EAA943C:1122334455667788 ADMIN:::AA9436C1D40CB53F3E7A20091C4B931C2F85252CC731BB25:8AC45ACDBD60F2FAD3081ECF005536EFA6009C21CA5FAF36:1122334455667788 ADMIN:::DCE867F0CB638DB2DBCC3576A52DC4612F85252CC731BB25:8990B33DAC65C5EF75073829894B911A983C1E260FBD1097:1122334455667788 ADMIN:::6F9D851D74C8A095C9DF672A1554BEBC2F85252CC731BB25:89953DE6F957B7DB5FE664D23AF3DE41DD38F5EC0A4A6EB0:1122334455667788 ADMIN:::CC96CC93B4DC9B7582273227FD61A5952F85252CC731BB25:76D3C3DEB0BB8EF1A1E41AB6A3F6C686A321CE016C624567:1122334455667788 ADMIN:::CC96CC93B4DC9B754DB66776827758D30B7892EEF2E3F2BC:DF58AE0F786BECC11BE11034DC53B21BDF1D73579AF868D1:1122334455667788 ADMIN:::DE5D1D85DAF6593D0A09FF32049013AB2F85252CC731BB25:526471D8C4A0ECC8AF05851804EA8FDD26848FA3CCC63152:1122334455667788 ADMIN:::B8489EDEE1058B43F3CE0F0ABE5A16872F85252CC731BB25:57B9C47A75335692F60E787E41CD16A292A21BC667B3FD02:1122334455667788 ADMIN:::2B6B134AF8D48F2A972BFF5660420D582F85252CC731BB25:5018402148E15A8D77CB22DD46F1449A2791416B73EE9C3D:1122334455667788 ADMIN:::BB49AEFD51ED0DCCD5BE291BD33BE3052F85252CC731BB25:C9B255750BD88AC72E03ADAFDA261E62618C943F7D59DAF5:1122334455667788 First, attack the "NETLM" "hashes" (case insensitive): host!solar:~/john/john-1.7.6-jumbo-4/run$ ./john --format=netlm pw-netntlm Loaded 15 password hashes with no different salts (LM C/R DES [netlm]) ADMIN (ADMIN) PASSWORD (ADMIN) 1234 (ADMIN) 123 (ADMIN) ASDFGH (ADMIN) 1 (ADMIN) 000000 (ADMIN) 00000000 (ADMIN) guesses: 8 time: 0:00:00:01 (3) c/s: 1306K trying: BETEMOR 12 (ADMIN) ROOT (ADMIN) guesses: 10 time: 0:00:00:04 (3) c/s: 1994K trying: MELACCT 00 (ADMIN) 0000 (ADMIN) guesses: 12 time: 0:00:00:07 (3) c/s: 1920K trying: KH6869 000 (ADMIN) 0000000 (ADMIN) guesses: 14 time: 0:00:00:19 (3) c/s: 1281K trying: CESKET1 Now let's try "NETNTLM" (case sensitive): host!solar:~/john/john-1.7.6-jumbo-4/run$ ./john --format=netntlm pw-netntlm Loaded 15 password hashes with no different salts (NTLMv1 C/R MD4 DES [netntlm]) ADMIN (ADMIN) password (ADMIN) 1234 (ADMIN) 123 (ADMIN) asdfgh (ADMIN) 1 (ADMIN) 000000 (ADMIN) 00000000 (ADMIN) guesses: 8 time: 0:00:00:01 (3) c/s: 1306K trying: sadie 12 (ADMIN) root (ADMIN) guesses: 10 time: 0:00:00:03 (3) c/s: 2371K trying: phdigh 0000 (ADMIN) 00 (ADMIN) guesses: 12 time: 0:00:00:06 (3) c/s: 2296K trying: rh3gap 000 (ADMIN) guesses: 13 time: 0:00:00:09 (3) c/s: 2033K trying: gte2g 0000000 (ADMIN) guesses: 14 time: 0:00:00:19 (3) c/s: 1626K trying: mbblum As you can see, either gets to 8 guesses in 1 second, and to 14 (out of 15 total) in under 19 seconds (the status line was displayed when I pressed a key; the actual guess occurred a bit earlier). It is also possible to go from known case insensitive passwords (cracked from NETLM hashes) to "crack the case" (from the NETNTLM hashes) nearly instantly, but this was not required in this case (we got to the same 14 hashes cracked quickly with a direct attack on NETNTLM as well). All of this was with JtR's default settings. Rainbow tables may be hot, but other approaches are viable as well, especially when the number of hashes or C/R's to audit is large (with rainbow tables, the attack time is per-hash, but with JtR the attack is against all hashes at once). I hope someone will find this helpful. Alexander
Powered by blists - more mailing lists
Confused about mailing lists and their use? Read about mailing lists on Wikipedia and check out these guidelines on proper formatting of your messages.