Source Code 1994 March/Source_Code_CD-ROM_Walnut_Creek_March_1994.iso/compsrcs/misc/volume28/crack/part05

Internet Message Format | 1992-03-18 | 32.8 KB

From: [e-mail] (Alec David Muffett)
Newsgroups: comp.sources.misc
Subject: v28i114: crack - The Unix Password Cracker, version 4.1, Part05/05
Message-ID: <[e-mail]>
Date: 18 Mar 92 14:39:33 GMT
Approved: [e-mail]
X-Md4-Signature: 40197e049e95f9f15ad937ab62008026

Submitted-by: [e-mail] (Alec David Muffett)
Posting-number: Volume 28, Issue 114
Archive-name: crack/part05
Environment: UNIX
Supersedes: crack: Volume 25, Issue 5-9

#! /bin/sh
# it by saving it into a file and typing "sh file".  To overwrite existing
# files, type "sh file -c".  You can also feed this as standard input via
# unshar, or by typing "sh <file", e.g..  If this archive is complete, you
# will see the following message at the end:
#        "End of archive 5 (of 5)."
# Contents:  Docs/readme.txt
# Wrapped by aem@aberfa on Wed Mar 18 14:08:39 1992
PATH=/bin:/usr/bin:/usr/ucb ; export PATH
if test -f 'Docs/readme.txt' -a "${1}" != "-c" ; then
  echo shar: Will not clobber existing file \"'Docs/readme.txt'\"
else
echo shar: Extracting \"'Docs/readme.txt'\" \(30990 characters\)
sed "s/^X//" >'Docs/readme.txt' <<'END_OF_FILE'
X
X
X
X
X
X
X
X
X
X                    "Crack Version 4.1"
X            A Sensible Password Checker for Unix
X
X
X                     Alec D.E. Muffett
X
X                   Unix Software Engineer
X                   Aberystwyth, Wales, UK
X      ([e-mail] or [e-mail])
X
X
X                          ABSTRACT
X
X          Crack is a freely available program  designed
X     to   find   standard   Unix   eight-character  DES
X     encrypted passwords  by  standard  guessing  tech-
X     niques outlined below.  It is written to be flexi-
X     ble, configurable and fast, and to be able to make
X     use  of  several  networked hosts via the Berkeley
X     rsh program (or similar), where possible.
X
X
X
X1.  Statement of Intent
X
XThis package is meant as a proving device to  aid  the  con-
Xstruction  of  secure  computer systems.  Users of Crack are
Xadvised that they may get severly hassled  by  authoritarian
Xtype sysadmin dudes if they run Crack without proper author-
Xisation.
X
X2.  Introduction to Version 4.0
X
XCrack is now into it's fourth version, and has been reworked
Xextensively  to provide extra functionality, and the purpose
Xof this release is to consolidate as much of this new  func-
Xtionality into as small a package as possible.  To this end,
XCrack may appear to be less configurable: it has been  writ-
Xten on the assumption that you run a fairly modern Unix, one
Xwith BSD functionality, and then patched in order to run  on
Xother systems.
X
XThis, surprisingly enough, has led to neater code,  and  has
Xmade  possible the introduction of greater flexibility which
Xsupercedes many of the options that could be  configured  in
Xearlier  versions  of  Crack.  In the same vein, some of the
Xolder options are now mandatory.  These,  such  as  feedback
Xmode  and  CRACK_PRINTOUT are no longer supported as options
Xand probably never will be again.  There is just  a  lot  of
Xwastage  in not running with them, and too many dependencies
Xin other functions to bother programming around them.
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 2 -
X
X
XThe user interface is basically identical  to  the  previous
Xversions,  although  some  people have asked about providing
XX-windows GUI's to Crack, I think it would  be  a  waste  of
Xtime to do so. Crack has far less options than your ordinary
Xversion of /bin/ls.
X
X3.  Introduction to Version 4.1
X
XVersion 4.1 of the Crack program is an attempt to extend the
Xfeatures  introduced  in v4.0 and provide hooks for external
Xlibraries such  as  Michael  Glad's  wonderful  UFC  crypt()
Ximplementation,  which (on some platforms) can outperform my
Xfcrypt() by a factor of 3.  I have also been  burdened  with
Xthe  task of making Crack's memory handling bombproof (hah!)
Xin the vague hope that it will survive running out of memory
Xon small machines.[1]
X
XThe extensions that I mention above regard the  addition  of
Xextra primitives to the dictionary processing language which
Xpermit the production of more concise dictionaries  contain-
Xing  words,  more  of which are likely to be passwords.  The
Xidea is to gain efficiency by removing  some  of  the  dross
Xfrom the generated dictionaries.
X
XCrack should (generally) be more  disk-space  efficient  now
Xthat  the  program  can  spot  dictionaries  which have been
Xcompressed using compress or pack and will  uncompress  them
Xon  the  fly  as  necessary  (using  zcat  or  pcat  respec-
Xtively).[2]
X
X4.  Crack Methodology - Part 1: Internals
X
XCrack takes as its input a  series  of  password  files  and
Xsource  dictionaries.  It merges the dictionaries, turns the
Xpassword files into a sorted list, and  generates  lists  of
Xpossible passwords from the merged dictionary or from infor-
Xmation gleaned about users from the password file.  It  does
X_________________________
X  [1] - or even on large ones. Brian Tompsett  at  Hull
Xtweaked  Crack  v3.3  until  it could run to completion
Xafter filling the swapspace on each  of  a  network  of
XSparcStation2's.  Due  to restructuring work on v4.0, I
Xhave had to  write  my  own  sorting  algorithm  &  re-
Ximplement  all of his tweaks from scratch, and can only
Xhope that I have emulated  the  bombproofness  of  this
Xdesirable (?) functionality.
X  [2] Note to people who are short on memory  or  swap:
Xdo  remember  that to do this Crack will have to fork()
X(via popen()) and might not be able to create  the  un-
Xcompressing  process.  Hence, if you intend to swaplock
Xyour machine, don't compress the  dictionaries.  Switch
Xthis off by editing the Crack shellscript.
X
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 3 -
X
X
Xnot attempt to remedy the problem of allowing users to  have
Xguessable  passwords,  and it should NOT be used in place of
Xgetting a really good, secure passwd program replacement.[3]
X
XThe above paragraphs define the purpose of Crack, and embody
Xa  great  deal of hard work, screams of Eureka!, drunkeness,
Xand a fair amount of swearing too.  There is a lot of think-
Xing, philosophy, and empirical guesswork behind the way that
XCrack attacks password files, and although it  is  not  per-
Xfect,  I  certainly hope that Crack will out-do most of it's
Xcompetitors.
X
XCrack works by making many individual passes over the  pass-
Xword  entries  that  you  supply to it.  Each pass generates
Xpassword guesses based upon a sequence of rules, supplied to
Xthe  program  by  the  user.  The  rules  are specified in a
Xsimplistic   language   in   the   files   gecos.rules   and
Xdicts.rules, to be found in the Scripts directory.  The dis-
Xtinction between these two files will be made clear later.
X
XThe rules are written as a simple string of characters, with
Xone  rule  to a line.  Blank lines, and comment lines begin-
Xning with a hash character #  are  ignored.   Trailing  whi-
Xtespace  is  also ignored.  The instructions in the rule are
Xfollowed from left to right, and are applied to the diction-
Xary  words one by one, as the words are loaded.  Some simple
Xpattern matching primitives are provided for selection  pur-
Xposes,  so  that  if  the dictionary word does not match the
Xpattern, it is ignored.  This  saves  on  time  and  memory.
XBefore  carrying  on,  I  suggest  that  you  browse through
XScripts/dicts.rules, take a look at the  rules  supplied  as
Xdefaults, and try to work out what they do.
X
XThe rules are stored in two different  files  for  two  dif-
Xferent  purposes.   Rules in Scripts/gecos.rules are applied
Xto data generated by Crack from the  pw_gecos  and  pw_gecos
Xentries  of  the user's password entry.  The data fed to the
Xgecos rules for the user aem, who  is  Alec  David  Muffett,
XSystems  would be: aem, Alec, David, Muffett, Systems, and a
Xseries of permutations of those  words,  either  re-ordering
Xthe  words  and  joining them together (eg: AlecMuffett), or
Xmaking up new words based on initial  letters  of  one  word
Xtaken with the rest of another (eg: AMuffett).[4]
X
XThe entire set of rules in gecos.rules is applied to each of
X_________________________
X  [3] See the end of ths document for more  information
Xabout passwd replacements.
X  [4] - and ASystems and DSystems, and MSystems, etc...
Xbecause  Crack  does  not  differentiate.  Hence,  care
Xshould be taken to check for redundancy when adding new
Xrules, so as not to waste time during the gecos pass.
X
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 4 -
X
X
Xthese words, which creates many more permutations and combi-
Xnations,  all  of which are tested.  Hence testing the pass-
Xword gecos information under Crack v4.0  and  upwards  takes
Xsomewhat   longer  than  previously,  but  it  is  far  more
Xthorough.
X
X
XAfter a pass has been made over  the  data  based  on  gecos
Xinformation,  Crack  makes  further passes over the password
Xdata using successive rules from the Scripts/dicts.rules  by
Xloading  the  whole  of Dicts/bigdict file into memory, with
Xthe rule being applied to each word from  that  file.   This
Xgenerates a resident dictionary, which is sorted and uniqued
Xso as to prevent wasting time on repetition. After each pass
Xis  completed, the memory used by the resident dictionary is
Xfreed up, and (hopefully) re-used when the  next  dictionary
Xis loaded.
X
XThe Dicts/bigdict dictionary is created by Crack by merging,
Xsorting,  and uniq'ing the source dictionaries, which are to
Xbe found in the directory DictSrc  and  which  may  also  be
Xnamed  in  the Crack shellscript, via the $STDDICT variable.
X(The default value of $STDDICT is /usr/dict/words).
X
XThe file DictSrc/bad_pws.dat is a dictionary which is  meant
Xto  provide  many  of  those common but non-dictionary pass-
Xwords, such as 12345678 or qwerty.
X
XIf you wish to provide a dictionary of your own,  just  copy
Xit  into  the  DictSrc  directory (use compress on it if you
Xwish to save space; Crack will unpack it  whilst  generating
Xthe  big  dictionary)  and  then  delete the contents of the
XDicts directory by running Scripts/spotless.  Your new  dic-
Xtionary will be merged in on the next run. For more informa-
Xtion on dictionary attacks, see the excellent  paper  called
X"Foiling  the  Cracker:  A  Survey  of, and Improvements to,
XPassword  Security"  by   Daniel   Klein,   available   from
Xftp.sei.cmu.edu  in  ~/pub/dvk/passwd.*.   Also, please read
Xthe APPENDIX file supplied with this distribution.[5]
X
XHaving described the method of cracking, perhaps  we  should
Xnow  investigate  the algorithm used to overlay the cracking
Xmechanism.
X
X
X
X_________________________
X  [5] Extra dictionaries (those detailed in Dan Klein's
Xpaper)   can   be   obtained  via  anonymous  FTP  from
Xftp.uu.net (137.39.1.9) as ~/pub/dictionaries.tar.Z; or
Xcheck  an Archie database for other possible sources of
Xdictionaries.
X
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 5 -
X
X
X5.  Crack Methodology - Part 2: Feedback Filters
X
XAs is stated above, Crack permutes  and  loads  dictionaries
Xdirectly   into  memory,  sorts  and  uniques  them,  before
Xattempting to use each of the words  as  a  guess  for  each
Xusers'  password.  If Crack correctly guesses a password, it
Xmarks the user as done and does not waste  further  time  on
Xtrying to break that users password.
X
XOnce Crack has finished a dictionary  pass,  it  sweeps  the
Xlist  of  users looking for the passwords it has cracked. It
Xstores the cracked passwords in both plaintext and encrypted
Xforms in a feedback file in the directory Runtime.  Feedback
Xfiles have names of the form Runtime/F*.
X
XThe purpose of this is so that, when Crack is next  invoked,
Xit  may recognise passwords that it has successfully cracked
Xbefore, and filter them  from  the  input  to  the  password
Xcracker.  This  provides  an instant list of crackable users
Xwho have not changed their passwords  since  the  last  time
XCrack was run. This list appears in a file with name out* in
Xthe $CRACK_OUT directory, or on stdout, if  foreground  mode
Xis invoked (see Options, below).
X
XIn a similar vein, when a Crack run terminates normally,  it
Xwrites out to the feedback file all encrypted passwords that
Xit has NOT succeeded in cracking.  Crack  will  then  ignore
Xall of these passwords next time you run it.
X
XObviously, this is not desirable if  you  frequently  change
Xyour  dictionaries  or  rules, and so there is a script pro-
Xvided,  Scripts/mrgfbk  which  sorts  your  feedback  files,
Xmerges  them  into one, and optionally removes all traces of
X'uncrackable' passwords, so that your  next  Crack  run  can
Xhave  a  go  at  passwords  it has not succeeded in breaking
Xbefore.
X
XMrgfbk is invoked automatically if you run Scripts/spotless.
X
X6.  Crack Methodology - Part 3: Execution and Networking
X
XEach time Crack is invoked, whether  networked  or  not,  it
Xgenerates  a diefile with a name of the form Runtime/D* (for
Xnetwork cracks, this file is generated by RCrack, and is  of
Xthe  form  Runtime/DR* which points to a real diefile, named
XRuntime/RD* - see below for details).
X
XThese diefiles contain debugging information about the  job,
Xand are generated so that all the jobs on the entire network
Xcan be called quickly by invoking Scripts/plaster.  Diefiles
Xdelete themselves after they have been run.
X
XAs you will read in the sections below, Crack has a -network
Xoption:   This   is  designed  to  be  a  simple  method  of
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 6 -
X
X
Xautomatically spreading the load of  password  cracking  out
Xover  several  machines on a network, preferably if they are
Xconnected by some form of networked filestore.
X
XWhen Crack -network is invoked, it filters its input in  the
Xordinary  way,  and  then splits its load up amongst several
Xmachines    which    are    specified    in     the     file
XScripts/network.conf.
X
XThis file contains a series  of  hostnames,  power  ratings,
Xflags,  etc,  relevant  to  the  running  of  Crack  on each
Xmachine.  Crack then calls Scripts/RCrack  to  use  the  rsh
Xcommand  (or  similar)  to  invoke Crack on the other hosts.
XSee the RCrack script, and the example network.conf file for
Xdetails.
X
X7.  Installation
X
XCrack is one of those most  unusual  of  beasties,  a  self-
Xinstalling  program.  Some people have complained about this
Xapparent weirdness, but it has  grown  up  with  Crack  ever
Xsince  the  earliest  network  version,  when I could not be
Xbothered to log into several different machines with several
Xdifferent   architectures,   just  in  order  to  build  the
Xbinaries. Once the necessary configuration options have been
Xset,  the  executables  are  created via make by running the
XCrack shellscript .
X
XCrack's configuration lies in two  files,  the  Crack  shell
Xscript,  which contains all the installation specific confi-
Xguration data, and the file Sources/conf.h,  which  contains
Xconfiguration options specific to various binary platforms.
X
XIn  the  Crack  shellscript,  you  will  have  to  edit  the
XCRACK_HOME  variable  to  the  correct value.  This variable
Xshould be set to an absolute path name  (names  relative  to
X~username  are  OK,  so  long  as you have some sort of csh)
Xthrough which the directory containing Crack may be accessed
Xon  ALL  the  machines that Crack will be run on. There is a
Xsimilar  variable  CRACK_OUT  which  specifies  where  Crack
Xshould  put  its output files - by default, this is the same
Xas $CRACK_HOME.
X
XYou will also have to edit the file Sources/conf.h and  work
Xout which switches to enable.  Each #define has a small note
Xexplaining its purpose.  Where I have been  in  doubt  about
Xthe portability of certain library functions, usually I have
Xre-written it, so you should be OK.  Let  me  know  of  your
Xproblems, if you have any.
X
XIf you will be using Crack -network you will  then  have  to
Xgenerate  a  Scripts/network.conf file. This contains a list
Xof hostnames to rsh to, what their binary  type  is  (useful
Xwhen   running   a   network   Crack  on  several  different
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 7 -
X
X
Xarchitectures), a guesstimate of their relative power  (take
Xyour  slowest machine as unary, and measure all others rela-
Xtive to it), and a list of per-host flags to  add  to  those
Xspecified on the Crack command line, when calling that host.
XThere is an example of such a file provided in  the  Scripts
Xdirectory - take a look at it.
X
XIf ever you wish to specify a more precise figure as to  the
Xrelative  power  of  your  machines,  or you are simply at a
Xloss, play with the command make tests in  the  source  code
Xdirectory.    This  can  provide  you  with  the  number  of
Xfcrypt()s that your machine can do per second,  which  is  a
Xnumber that you can plug into your network.conf as a measure
Xof your machines' power (after  rounding  the  value  to  an
Xinteger).
X
X8.  Usage
X
XOkay, so, let's assume  that  you  have  edited  your  Crack
Xscript,  and  your Sources/conf.h file, where do you go from
Xhere ?
X
X
XCrack [options] [bindir] /etc/passwd [...other passwd files]
X
XCrack -network [options] /etc/passwd [...other passwd files]
X
X
XWhere bindir is the optional name of the directory where you
Xwant  the binaries installed.  This is useful where you want
Xto be able to run versions of  Crack  on  several  different
Xarchitectures.  If  bindir does not exist, a warning will be
Xissued, and the directory created.
X
X     Note: bindir defaults to the name generic  if  not
X     supplied.
X
X
XNotes for  Yellow  Pages  (NIS)  Users:  I  have  occasional
Xqueries  about  how  to get Crack running from a YP password
Xfile.  There are several methods, but by far the simplest is
Xto generate a passwd format file by running:-
X
X                  ypcat passwd > passwd.yp
X
Xand then running Crack on this file.
X
X9.  Options
X
X-f   Runs Crack in foreground mode, ie: the password cracker
X     is  not backgrounded, and messages appear on stdout and
X     stderr as you would expect.  This option is only really
X     useful  for very small password files, or when you want
X     to put a wrapper script around Crack.
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 8 -
X
X
X     Foreground mode is disabled if you  try  running  Crack
X     -network  -f on the command line, because of the insen-
X     sibility of rshing to several machines in turn, waiting
X     for  each  one  to finish before calling the next. How-
X     ever, please read the section  about  Network  Cracking
X     without NFS/RFS, below.
X
X-v   Sets verbose mode, whereby Crack will print every guess
X     it is trying on a per-user basis.  This is a very quick
X     way of flooding your filestore, but useful if you think
X     something is going wrong.
X
X-m   Sends mail to any user  whose  password  you  crack  by
X     invoking  Scripts/nastygram  with  their username as an
X     argument.  The reason for using the script is so that a
X     degree of flexibility in the format of the mail message
X     is supplied; ie: you don't have to  recompile  code  in
X     order to change the message.[6]
X
X-nvalue
X     Sets the process to be nice()ed to value, so, for exam-
X     ple,  the  switch -n19 sets the Crack process to run at
X     the lowest priority.
X
X-network
X     Throws Crack into network mode, in which it  reads  the
X     Scripts/network.conf file, splits its input into chunks
X     which are sized according to the power  of  the  target
X     machine,  and  calls  rsh to run Crack on that machine.
X     Options for Crack running on the target machine may  be
X     supplied  on  the  command line (eg: verbose or recover
X     mode), or in the network.conf file if they  pertain  to
X     specific hosts (eg: nice() values).
X
X-r<pointfile>
X     This is only for use  when  running  in  recover  mode.
X     When  a  running  Crack  starts pass 2, it periodically
X     saves its state in a pointfile, with a name of the form
X     Runtime/P.*  This file can be used to recover where you
X     were should a  host  crash.   Simply  invoke  Crack  in
X     exactly  the  same  manner  as  the last time, with the
X     addition of the -r switch,  (eg:  -rRuntime/Pfred12345)
X_________________________
X  [6] I'm uncertain about the wisdom of mailing someone
Xlike this.  If someone browses your cracked user's mail
Xsomehow, it's like a great big neon  sign  pointing  at
Xthe  user  saying "This Is A Crackable Account - Go For
XIt!".  Not to mention the false sense  of  security  it
Xengenders  in  the  System Manager that he's "informed"
Xthe user to change his  password.   What  if  the  user
Xdoesn't  log  on  for 3 months? However, so many people
Xhave wired it into their own versions of Crack, I  sup-
Xpose it must be provided... AEM
X<1b>9
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 9 -
X
X
X     switch.  Crack will startup and read the file, and jump
X     to  roughly  where  it left off.  If you are cracking a
X     very large password file, this can save oodles of  time
X     after a crash.
X
X     If you were running a network Crack, then the jobs will
X     again  be spawned onto all the machines of the original
X     Crack.  The program will then check that the host it is
X     running   on  is  the  same  as  is  mentioned  in  the
X     pointfile.  If it is not, it will quietly  die.   Thus,
X     assuming that you supply the same input data and do not
X     change your network.conf file,  Crack  should  pick  up
X     where  it  left off.  This is a bit inelegant, but it's
X     better than nothing at the moment.
X
X     The method of  error  recovery  outlined  above  causes
X     headaches  for  users who want to do multiprocessing on
X     parallel architectures.  Crack is in no  way  parallel,
X     and  because  of the way it's structured (reading stdin
X     from shellscript frontends) it is a pain to divide  the
X     work amongst several processes via fork()ing.
X
X     The hack solution to get several copies of  Crack  run-
X     ning  on one machine with n processors at the moment is
X     to insert n copies  of  the  entry  for  your  parallel
X     machine  into the Scripts/network.conf file. If you use
X     the -r option in these circumstances however, you  will
X     get n copies of the recovered process running, only one
X     of them will have the correct input data.
X
X     The old solution to this problem (see old documentation
X     if  you  are interested) has been negated by the intro-
X     duction of feedback mode, so the best bet in this  par-
X     ticular  situation  is to wait until the other jobs are
X     done (and have written out lists of  uncrackable  pass-
X     words),  and then re-start the jobs from scratch.  Any-
X     one whose password was not cracked  on  the  first  run
X     will be ignored on the second, if they have not changed
X     it since.  This is inelegant, but it's the best  I  can
X     do in the limited time available.
X
X10.  Support Scripts
X
XThe Scripts directory contains a small number of support and
Xutility  scripts,  some  of which are designed to help Crack
Xusers check their progress.  Briefly, the most  useful  ones
Xare:-
X
XScripts/shadmrg
X     This is a small (but  hopefully  readable)  script  for
X     merging  /etc/passwd  and /etc/shadow on System V style
X     shadow password systems.  It produces the  merged  data
X     to stdout, and will need redirecting into a file before
X     Crack can work on it.  The script is meant to be fairly
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 10 -
X
X
X     lucid,  on the grounds that I worry that there are many
X     shadowing schemes out there, and perhaps not  all  have
X     the same data format.
X
X     I have not wired this facility into the  Crack  command
X     itself because the world does NOT revolve around System
X     V yet, regardless of what some  people  would  have  me
X     believe,  and I believe that the lack of direct support
X     for NIS outlined above, sets a  precedent.   There  are
X     just  too  many  incompatibilities  in  shadow password
X     schemes for me to hardwire anything.
X
XScripts/plaster
X     which is named after a dumb joke, but is a simple fron-
X     tend  to  the Runtime/D* diefiles that each copy of the
X     password cracker  generates.  Invoking  Scripts/plaster
X     will  kill  off  all copies of the password cracker you
X     are running, over the network or otherwise.
X
XScripts/status
X     This script rshes to  each  machine  mentioned  in  the
X     Scripts/network.conf  file,  and provides some informa-
X     tion about processes and uptime on that machine.   This
X     is  useful when you want to find out just how well your
X     password crackers are getting on during a  Crack  -net-
X     work.
X
XScripts/{clean,spotless}
X     These are really just frontends to a makefile. Invoking
X     Scripts/clean  tidies  up the Crack home directory, and
X     removes probably unwanted files, but  leaves  the  pre-
X     processed  dictionary bigdict intact.  Scripts/spotless
X     does the same as Scripts/clean but obliterates  bigdict
X     and  old  output files too, and compresses the feedback
X     files into one.
X
XScripts/nastygram
X     This is the shellscript that is invoked by the password
X     cracker  to send mail to users who have guessable pass-
X     words, if the -m option is used. Edit it at  your  lei-
X     sure to suit your system.
X
XScripts/guess2fbk
X     This script takes your  out*  files  as  arguments  and
X     reformats  the  'Guessed'  lines  into a slightly messy
X     feedback file, suitable for storing with the others.
X
X     An occasion where this might be  useful  is  when  your
X     cracker  has  guessed  many peoples passwords, and then
X     died for some reason (a crash?) before writing out  the
X     guesses to a feedback file.  Running
X
X             Scripts/guess2fbk out* >> Runtime/F.new
X
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 11 -
X
X
X     will save the work that has been done.
X
X11.  Network Cracking without NFS/RFS
X
XFor those users who have some form of rsh  command,  but  do
Xnot  have  a  a  networked  filestore running between hosts,
Xthere is now a solution which will allow you to do networked
Xcracking, proposed to me by Brian Tompsett at Hull.  Person-
Xally, I consider the idea to be potty, but it fills in miss-
Xing functionality in a wonderfully tacky manner.
X
XFrom the documentation above, you will note that Crack  will
Xundo  the -f (output in foreground) option, if it is invoked
Xwith the -network switch at the same time (see  the  Options
Xsection  above).  This is true, but it does not apply if you
Xspecify -f option in the network.conf file.
X
XThe practical upshot of doing this is that remote copies  of
XCrack  can  be  made  to read from stdin and write to stdout
Xover a network link, and thus remote  processing  is  accom-
Xplished.   I  have  tweaked  Crack in such a way, therefore,
Xthat if the -f option is specified amongst  the  crack-flags
Xof  a  host  in  the network.conf, rather than backgrounding
Xitself on the remote host, the rsh command on the server  is
Xbackgrounded, and output is written directly to the files on
Xthe server's filestore.
X
XThere are restrictions upon this  method,  mostly  involving
Xthe number of processes that a user may run on the server at
Xany one time, and that you will  have  to  collect  feedback
Xoutput  together  manually  (dropping  it  into  the Runtime
Xdirectory on the server).  However, it works. Also,  if  you
Xtry  to use rsh as another user, you will suffer problems if
Xrsh insists on reading something from your terminal  (eg:  a
Xpassword  for  the  remote account).  Also, recovering using
Xcheckpointing goes out the window  unless  you  specify  the
Xname of the pointfile as it is named on the remote machine.
X
X12.  UFC Support and notes on fast crypt() implementations
X
XThe stdlib version of the crypt() subroutine  is  incredibly
Xslow.   It is a massive bottleneck to the execution of Crack
Xand on typical platforms that you get at universities, it is
Xrare to find a machine which will achieve more than 50 stan-
Xdard crypt() s per second.   On  low-end  diskless  worksta-
Xtions,  you may expect 2 or 3 per second.  It was this slow-
Xness of the crypt() algorithm which originally supplied much
Xof the security Unix needed.[7]
X
X_________________________
X  [7] See: "Password Security, A Case History"  by  Bob
XMorris & Ken Thomson, in the Unix Programmer Docs.
X
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 12 -
X
X
XThere are now many implementations  of  faster  versions  of
Xcrypt()  to  be found on the network.  The one supplied with
XCrack v3.2 and upwards is called fcrypt().   It  was  origi-
Xnally written in May 1986 by Robert Baldwin at MIT, and is a
Xgood version of the crypt() subroutine.  I received  a  copy
Xfrom Icarus Sparry at Bath University, who had made a couple
Xof portability enhancements to the code.
X
XI rewrote most of the tables and the KeySchedule  generating
Xalgorithm  in  the original fdes-init.c to knock 40% off the
Xexecution overhead of fcrypt()  in  the  form  that  it  was
Xshipped  to  me.   I inlined a bunch of stuff, put it into a
Xsingle file, got  some  advice  from  Matt  Bishop  and  Bob
XBaldwin  [both  of whom I am greatly indebted to] about what
Xto do to the xform() routine  and  to  the  fcrypt  function
Xitself,  and  tidied  up some algorithms.  I have also added
Xmore lookup tables and reduced several  formula  for  faster
Xuse.   Fcrypt() is now barely recognisable as being based on
Xits former incarnation, and it is 3x faster.
X
XOn a DecStation 5000/200, fcrypt() is about 16 times  faster
Xthan  the  standard  crypt (your mileage may vary with other
Xarchitectures and compilers).  This speed puts fcrypt() into
Xthe "moderately fast" league of crypt implementations.
X
XAmongst other crypt  implementations  available  is  UFC  by
XMichael  Glad.   UFC-crypt is a version of the crypt subrou-
Xtine which  is  optimised  for  machines  with  32-bit  long
Xintegers  and  generally outperforms my fcrypt() by a factor
Xof between 1 and 3, for a tradeoff of  large  memory  usage,
Xand memory-cache unfriendliness.  Hooks for even more optim-
Xised assembler versions of crypt()  are  also  provided  for
Xsome  platforms  (Sun,  HP, ...).  Getting UFC to work on 16
Xbit architectures is nearly impossible.
X
XHowever, on most architectures,  UFC  generates  a  stunning
Xincrease  in  the power of Crack, and so, from v4.1 onwards,
XCrack is written to automatically make use of UFC if it  can
Xfind  it.   All  that you have to do is to obtain a suitable
Xcopy of UFC (preferably a version which mentions that it  is
Xcompatible  with  Crack v4.1, and unpack it into a directory
Xcalled ufc-crypt in $CRACK_HOME, and then  delete  your  old
Xbinaries.   UFC  will then be detected, compiled, tested and
Xused in preference to fcrypt() by the Crack  program,  wher-
Xever possible.
X
X13.  Conclusions
X
XWhat can be done about brute force attacks on your  password
Xfile ?
X
XYou must get  a  drop-in  replacement  for  the  passwd  and
Xyppasswd  commands; one which will stop people from choosing
Xbad  passwords  in  the  first  place.   There  are  several
X
X
X
X                       March 3, 1992
X
X
X
X
X
X                           - 13 -
X
X
Xprograms  to  do  this;  Matt  Bishop's  passwd+  and  Clyde
XHoover's npasswd program are good examples which are  freely
Xavailable.   Consult  an Archie database for more details on
Xwhere you can get them from.
X
XIt would be nice if an organisation (such as CERT?) could be
Xpersuaded  to  supply  skeletons of sensible passwd commands
Xfor the public good, as  well  as  an  archive  of  security
Xrelated utilities[8] on top of the excellent COPS.  However,
Xfor Unix security to improve on a global scale, we will also
Xrequire pressure on the vendors, so that programs are  writ-
Xten correctly from the beginning.
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X_________________________
X  [8] COPS  is  available  for   anonymous   FTP   from
Xcert.sei.cmu.edu (128.237.253.5) in ~/cops
X
X
X
X
X                       March 3, 1992
X
X
END_OF_FILE
if test 30990 -ne `wc -c <'Docs/readme.txt'`; then
    echo shar: \"'Docs/readme.txt'\" unpacked with wrong size!
fi
# end of 'Docs/readme.txt'
fi
echo shar: End of archive 5 \(of 5\).
cp /dev/null ark5isdone
MISSING=""
for I in 1 2 3 4 5 ; do
    if test ! -f ark${I}isdone ; then
    MISSING="${MISSING} ${I}"
    fi
done
if test "${MISSING}" = "" ; then
    echo You have unpacked all 5 archives.
    rm -f ark[1-9]isdone
else
    echo You still need to unpack the following archives:
    echo "        " ${MISSING}
fi
##  End of shell archive.
exit 0
exit 0 # Just in case...