Skip to content

Instantly share code, notes, and snippets.

@dodyw
Last active October 11, 2019 10:30
Show Gist options
  • Star 13 You must be signed in to star a gist
  • Fork 5 You must be signed in to fork a gist
  • Save dodyw/8205047 to your computer and use it in GitHub Desktop.
Save dodyw/8205047 to your computer and use it in GitHub Desktop.
Perl script to find suspicious scripts. This script is taken from http://cbl.abuseat.org/findbot.pl
#!/usr/bin/perl
# The above line may need to be changed to point at your version of Perl
#
# This script attempts to find malicious files/scripts on your machine.
# It specifically looks for spambots that we're aware of, as well
# as "suspicious" constructs in various scripting languages.
#
# Normally it should be run as root.
#
# By default, findbot.pl scans the directories /tmp, /usr/tmp, /home and
# /var/www. This script isn't fast. So if you know where to look you can
# speed things up by giving just the directories that you suspect has the
# malware.
#
# You can often find out what user is infected by using:
# lsof -i | grep smtp
# and looking for processes that are NOT your mail server.
#
# If you're successful finding the user, you need to look everywhere the user
# has write permissions - and you can run findbot.pl faster, by something like:
#
# findbot.pl /tmp /usr/tmp /home/<user> <user's web directory>
#
# There are two types of "detections" - "suspicious files" are files that contain
# things that -may- be malicious.
# "malware" is definitely malicious software.
#
# This script needs the following command line utilities. It will not run
# if it can't find them, you will have to install them yourself:
# - "md5sum" (Linux) or "md5" (FreeBSD etc) this appears to be standard
# core utilities.
# - "strings" - on Linux this is in the "binutils" package
# - "file" - on Linux this is in the "file" package.
#
# Usage:
# findbot.pl [directories...]
#
# If a list of directories is supplied, it's used, otherwise,
# /tmp, /usr/tmp, /home and /var/www are use by default.
#
# Very simple web malware detection module.
# Version 0.02 2013/01/02 Ray
# .01 -> .02:
# - more strings of bad software
# - search for encoded perl scripts
# .02 -> .03: 2013/01/10 Ray
# - speed up
# - MD5 stuff
# .03 -> .04: 2013/01/13 Ray
# - improved docs
# .04 -> .005: 2013/01/20 Ray
# - more patterns
# - MAXLINES way too small
my $access = '(\.htaccess)';
my $accesspat = '(RewriteRule)';
my $scripts = '\.(php|pl|cgi|bak)$';
my $scriptpat = '(r57|c99|web shell|passthru|shell_exec|phpinfo|base64_decode|edoced_46esab|PHPShell|EHLO|MAIL FROM|RCPT TO|fsockopen|\$random_num\.qmail|getmxrr)';
my @defaultdirs = ('/tmp', '/usr/tmp', '/home', '/var/www');
my $MAXLINES = 400;
my($strings, $md5sum, $file, %badhash);
&inithelpers;
&badhashes;
#my $executable = '^(sshd|cache|exim|sh|bash)$';
if ($ARGV[0] =~ /^-/) {
my $l = join(',', @defaultdirs);
print STDERR <<EOF;
usage: $0 [directories to scan...]
If no directories specified, script uses:
$l
EOF
exit 0;
}
if (!scalar(@ARGV)) {
push(@ARGV, @defaultdirs);
}
for my $dir (@ARGV) {
&recursion($dir);
}
sub recursion {
my ($dir) = @_;
my (@list);
if (!opendir(I, "$dir")) {
return if $! =~ /no such file/i;
print STDERR "$dir: Can't open: $!, skipping\n";
return;
}
@list = readdir(I);
closedir(I);
for my $mfile (@list) {
next if $mfile =~ /^\.\.?$/; # skip . and ..
my $cf = $currentfile = "$dir/$mfile";
$cf =~ s/'/'"'"'/g; # hide single-quotes in filename
$cf = "'$cf'"; # bury in single-quotes
if (-d $currentfile && ! -l $currentfile) {
&recursion($currentfile); # don't scan symlinks
next;
}
next if ! -f $currentfile;
if ($mfile =~ /$scripts/) {
&scanfile($currentfile, $scriptpat);
} elsif ($mfile =~ /$access/) {
&scanfile($currentfile, $accesspat);
}
# up to here it's fast.
next if -s $currentfile > 1000000 || -s $currentfile < 2000;
#print STDERR "$currentfile\n";
my $type = `$file $cf`;
if ($type =~ /(ELF|\d\d-bit).*executable/ || $currentfile =~ /\.(exe|scr|com)$/) {
#print STDERR "cf: $cf\n";
my $checksum = `$md5sum $cf`;
chomp($checksum);
$checksum =~ s/\s.*//;
if ($badhash{$checksum}) {
print STDERR "$currentfile: Malware detected!\n";
next;
}
my $strings = `$strings $cf`;
if ($strings =~ /\/usr\/bin\/perl/sm) {
print STDERR "$currentfile: possible binary-encoded-perl\n";
next;
}
}
}
}
sub scanfile {
my ($currentfile, $patterns) = @_;
#print $currentfile, "\n";
open(I, "<$currentfile") || next;
my $linecount = 1;
while(<I>) {
chomp;
if ($_ =~ /$patterns/) {
my $pat = $1;
my $string = $_;
if ($string =~ /^(.*)$pat(.*)$/) {
$string = substr($1, length($1)-10, 10) .
$pat .
substr($2, 0, 10);
}
#$string =~ s/^.*(.{,10}$pat.{,10}).*$/... $1 .../;
print "$currentfile: Suspicious($pat): $string\n";
last;
}
last if $linecount++ > $MAXLINES;
}
close(I);
}
sub inithelpers {
if (-x '/usr/bin/md5sum') {
$md5sum = '/usr/bin/md5sum';
} elsif (-x '/sbin/md5') {
$md5sum = '/sbin/md5 -q';
}
for my $x (('/bin', '/usr/bin')) {
if (-x "$x/strings") {
$strings = "$x/strings";
}
if (-x "$x/file") {
$file = "$x/file";
}
}
die "Can't find md5 checksumming tool - normally in Linux coretools package" if !$md5sum;
die "Can't find strings tool - normally in Linux bintools package" if !$strings;
die "Can't find file tool - normally in file package" if !$file;
}
sub badhashes {
map { $badhash{$_} = 1; } ((
'f7536bb412d6c4573fd6fd819e1b07bb',
'0fdb34f48166dae57ff410d723efd3f7',
'396d1fb94d79b732f6ab2fa6c5f3ed39',
'fd3c01133946d59ace4fdb49dde93268', #Directmailer .exe Windows binary
));
}
@kariha
Copy link

kariha commented Sep 29, 2015

i want a log file. how can i do?

@raytrax
Copy link

raytrax commented Dec 11, 2017

Yes... easy way: ./findbot.pl > Suspicious.txt

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment