#!/bin/bash
## Originally...
# Copyright (c) 2004 Matthias S. Benkmann <article AT winterdrache DOT de>
# You may do everything with this code except misrepresent its origin.
# PROVIDED `AS IS' WITH ABSOLUTELY NO WARRANTY OF ANY KIND!

# Copyright (C) 2014 Steve Youngs <steve@steveyoungs.com>
# many updates/tweaks --SY.

# The following list should contain the mount points of all filesystems
# that are to be scanned as a space-separated list within parentheses. 
# / will usually be in this list and if you have /usr
# on a separate partition, it will also be in this list. Other non-special
# filesystems where suspicious files could be located should also be put in 
# this list.
# Mount points whose filesystems are special, such as procfs or sysfs should
# not be in this list. 

fs_to_scan=(/)

## Bastard settings
#fs_to_scan=(\
#	/ \
#	/opt \
#	/usr \
#	/usr/src \
#	/var)

# Files with a path prefix found in the following list are ignored.
# DO !!!!NOT!!! PUT /usr/src OR WHATEVER THE HOME DIRECTORY prefix is for your
# package users into this list!!! You DO want to scan those directories in
# order to spot e.g. world-writable tarballs and other abominations that
# may have crept in.
# Ideally, this list should be empty.

prune_prefixes=(/root /{,*/{,*/}}lost+found) #NO TRAILING SLASHES!!!

## Bastard settings
#prune_prefixes=(\
#	/{,*/{,*/}}lost+found \
#       /root \
#	/opt/pgsql/data \
#	/opt/sql-ledger/{spool,templates,users,css} \
#       /etc/apache/ssl.key \
#	/etc/audisp/plugins.d \
#       /etc/cups/ssl \
#       /etc/firewall \
#	/etc/polkit-1/rules.d \
#       /etc/skel \
#       /etc/ssl/private \
#	/etc/sudoers.d \
#	/var/lib/{sasl,sudo,net-snmp,udisks{,2},NetworkManager} \
#	/var/log \
#	/usr/lib/pkgusr \
#	/usr/share/polkit-1/rules.d \
#	/var/tmp \
#       /var/{cache,chroot,run,snmp,spool} \
#       /var/lib/{sshd,nfs,spamassassin,pulse}) #NO TRAILING SLASHES!!!!

# Set the following to `-noleaf' if you are scanning non-UNIX filesystems
# like MS-DOS, CD-ROM etc.  But only do so if you really need it as it
# will slow the search significantly.
# NOLEAF='-noleaf'
NOLEAF=

# If the following variable is set to "yes", then files that contain
# control characters or other non-printable characters (except for space)
# will be reported as suspicious. 
# This test slows down the search considerably!
#enable_illchars=yes
enable_illchars=no


# suppress ugly debug output from shell
trap ':' SIGPIPE

# "-false" as 1st argument is used when called by list_suspicious_files_from
if [ $# -ge 1 -a "$1" != "-false" ]; then
    echo 1>&2 
    echo 1>&2 "USAGE: ${0##*/}"
    echo 1>&2 
    echo 1>&2 '  Outputs a categorized list of files and directories with properties'
    echo 1>&2 '  that could mean trouble and should be investigated.'
    echo 1>&2
    exit 1
fi


usergroupmatch=(-true)
if [ "$1" = "-false" ]; then
    usergroupmatch=(\( "$@" \))
fi

# construct find commands that match the prune_prefixes. Each prefix will be
# matched as -path <prefix> -or -path <prefix>/*
# so that the directory itself and all subdirectories are matched.
y=(\( -false)
for ((i=0; $i<${#prune_prefixes[@]}; i=$i+1)) 
do
    y[${#y[@]}]='-or'
    y[${#y[@]}]=-path
    y[${#y[@]}]="${prune_prefixes[$i]}"
    y[${#y[@]}]='-or'
    y[${#y[@]}]=-path
    y[${#y[@]}]="${prune_prefixes[$i]}/*"
done
y[${#y[@]}]=')'

illchars=( $'\x1'  $'\x2'  $'\x3'  $'\x4'  $'\x5'  $'\x6'  $'\x7'  $'\x8'  
  $'\x9'  $'\xA'  $'\xB'  $'\xC'  $'\xD'  $'\xE'  $'\xF'  $'\x10'  $'\x11' 
  $'\x12'  $'\x13'  $'\x14'  $'\x15'  $'\x16'  $'\x17'  $'\x18'  $'\x19'  
  $'\x1A'  $'\x1B'  $'\x1C'  $'\x1D'  $'\x1E'  $'\x1F'  $'\x7f' $'\x80' 
  $'\x81'  $'\x82'  $'\x83'  $'\x84'  $'\x85'  $'\x86'  $'\x87'  $'\x88'  
  $'\x89'  $'\x8A'  $'\x8B'  $'\x8C'  $'\x8D'  $'\x8E'  $'\x8F'  $'\x90'  
  $'\x91'  $'\x92'  $'\x93'  $'\x94'  $'\x95'  $'\x96'  $'\x97'  $'\x98'  
  $'\x99'  $'\x9A'  $'\x9B'  $'\x9C'  $'\x9D'  $'\x9E'  $'\x9F' ) 


if [ "$enable_illchars" = yes ]; then

    illname=(\( -false)
    for ((i=0; $i<${#illchars[@]}; i=$i+1)) 
    do
        #handle bash \x7f error
        if [ "*${illchars[$i]}*" = "**" ]; then
	    illchars[$i]=$'\x80'  #'
	fi
	illname[${#illname[@]}]='-or'
	illname[${#illname[@]}]=-name
        illname[${#illname[@]}]="*${illchars[$i]}*"
    done
    illname[${#illname[@]}]=')'

    illlink=(\( -false)
    for ((i=0; $i<${#illchars[@]}; i=$i+1)) 
    do
        illlink[${#illlink[@]}]='-or'
        illlink[${#illlink[@]}]=-lname
        illlink[${#illlink[@]}]="*${illchars[$i]}*"
    done
    illlink[${#illlink[@]}]=')'
else #if [ "$enable_illchars" = no ]
    illlink=(-false)
    illname=(-false)
fi

# $1=section heading
# $2=inode message
report()
{
    echo -printf "increment_code_here"
    echo -printf 
    echo "1 ${1}\\n" | sed 's/ /\\040/g'
    echo -printf "insert_code_here"
  
    if [ -n "$2" ]; then
	echo -printf 
	echo "2 %i 1 ${2}\\n" | sed 's/ /\\040/g'
	echo -printf "insert_code_here"
	echo -printf 
	echo "2 %i 2 " | sed 's/ /\\040/g'
    else
	echo -printf "2\\040" 
    fi
  
    echo -exec ls -T 0 -ladQ {} \;
}


filegoodperm=(\( -perm 644 -or -perm 755 -or -perm 555 -or -perm 444 -or -perm 600 -or -perm 700 -or -perm 640 \))
dirgoodperm=(\( -perm 755 -or -perm 555 -or -perm 700 -or -perm 750 \))

good=( \( 
          -not \( -not -type d -links +1 \)
          -not -nouser -not -nogroup 
          -not \( "${illname[@]}" \) 
          -not \( "${illlink[@]}" \) 
       \) 
       -and
\(
      \( -type f -not -group install "${filegoodperm[@]}" \) 
  -or \( -type d -not -group install "${dirgoodperm[@]}" \)
  -or \( -type d -group install \( -perm 1775 \) \)
  -or \( -type d -group root -user root -path "/tmp" \( -perm 1777 \) \)
  -or \( -type d -group root -user root -path "/var/tmp" \( -perm 1777 \) \)
  -or \( -not -type d -not -type f -not -type l -path "/dev/*" \)
  -or \( -type l \( -xtype b -or -xtype c -or -xtype d -or -xtype p -or -xtype f \) \)
\)
)

bad=(
    \( "${illname[@]}" $(report  "NON-PRINTABLE CHAR IN NAME")  \)
 OP \( "${illlink[@]}" $(report  "NON-PRINTABLE-CHAR IN LINK-TARGET")  \)
 OP \( -type f -perm -4000 $(report  "SETUID FILES")  \)
 OP \( -type f -perm -2000 $(report  "SETGID FILES")  \)
 OP \( -type f -perm -1000 $(report  "STICKY FILES")  \)
 OP \( -type d -perm -2000 $(report  "GROUP-KEEPING DIRECTORIES")  \)
 OP \( -type d -not -group install -perm -1000  $(report  "STICKY DIRECTORIES")  \)
 OP \( -type f -perm -g+w  $(report  "GROUP-WRITABLE FILES")  \)
 OP \( -type f -perm -o+w  $(report  "WORLD-WRITABLE FILES")  \)
 OP \( -type d -perm -g+w  $(report  "GROUP-WRITABLE DIRECTORIES")  \)
 OP \( -type d -perm -o+w  $(report  "WORLD-WRITABLE DIRECTORIES")  \)
 OP \( -not \( -type f -or -type l -or -type d \) -not -path "/dev/*"  $(report  "SPECIAL FILES OUTSIDE /dev")  \)
 OP \( -type d -group install -not -perm 1755  $(report  "INSTALL DIRECTORIES WITH UNUSUAL PERMISSIONS")  \)
 OP \( -type f -group install $(report  "FILES ASSIGNED TO GROUP INSTALL")  \)
 OP \( -type l -not \( -xtype b -or -xtype c -or -xtype d -or -xtype p -or -xtype f \) $(report  "SYMLINKS POSSIBLY BROKEN OR LOOP")  \)
 OP \( -not -type d -links +1 $(report "HARDLINKED FILES" "Inode %i is shared by %n files, including") \)
 OP \( -nouser  $(report  "THINGS HAVING UID WITH NO ASSIGNED USER NAME")  \)
 OP \( -nogroup  $(report  "THINGS HAVING GID WITH NO ASSIGNED GROUP NAME")  \)
 OP \( -type f -not -group install -not "${filegoodperm[@]}"  $(report  "FILES WITH UNUSUAL PERMISSIONS")  \)
 OP \( -type d -not -group install -not "${dirgoodperm[@]}"  $(report  "DIRECTORIES WITH UNUSUAL PERMISSIONS")  \)
)

# insert unique codes for the messages
code=100
for ((i=0; $i<${#bad[@]}; i=$i+1)) 
do
    if [ "${bad[$i]}" = "increment_code_here" ]; then
	code=$(($code + 1))
	bad[$i]=$code
    elif [ "${bad[$i]}" = "insert_code_here" ]; then
	bad[$i]=$code
    fi
done

allbad=()  #all bad matches are reported
onebad=()  #only the first bad match is reported
for ((i=0; $i<${#bad[@]}; i=$i+1)) 
do
    if [ "${bad[$i]}" = "OP" ]; then
	allbad[$i]=","
	onebad[$i]="-or"
    else
	allbad[$i]="${bad[$i]}"
	onebad[$i]="${bad[$i]}"
    fi
done

# Add a default case to onebad.
# This should never be hit, because the explicit cases should catch all
# files, but just in case I've missed something, this will catch it.
onebad=("${onebad[@]}" -or  $(report  "WEIRD SHIT") )

# make allbad always return false
allbad=("${allbad[@]}" , -false)

cmd=( "${usergroupmatch[@]}" -and
     \( \( "${good[@]}" \) -or \( "${allbad[@]}" \) -or \( "${onebad[@]}" \) \)
    )

# In the following find command, the part
# -not ( ( "${y[@]}" -prune ) -or "${y[@]}" )
# is responsible for preventing the files that match prune_prefixes from
# being processed. The 2nd "${y[@]}" may seem redundant, but it isn't, because
# -prune has no effect and is always false when -depth is used.
find "${fs_to_scan[@]}" -xdev $NOLEAF \
    -not \( \( "${y[@]}" -prune \) -or "${y[@]}" \) \
    -and \( "${cmd[@]}" \) 2>/dev/null | 
sed 's/^\(...2\) \([0-9]\+ 2 \)\?\([^ ]\+\) \+[^ ]\+ \+\([^ ]\+\) \+\([^ ]\+\) \+[^"]\+\(".\+\)/\1 \2\3 \6  \4:\5/' |
sort -u | 
sed 's/^...1 /\'$'\n''/;s/^...2 [0-9]\+ 1 /\'$'\n''  /;s/^...2 [0-9]\+ 2 /    /;s/^...2 /  /'