Eu tenho um arquivo de regras. Em cada linha há um sid: number com um rev: number.
Eu quero comparar este arquivo com um arquivo atualizado, mas nem todas as linhas serão atualizadas.
Se uma das linhas com o mesmo sid: number tiver uma maior rev: number eu preciso substituí-la pela maior rev: number.
Este é o tipo de lugar onde estou até agora
grep -oP "sid:[0-9]{0,11}; rev:[0-9]{0,3}" all_rules.rules |
while read line; do
if grep -q "$line" /home/path/update_rules.rules; then
echo updated;
else
echo > /dev/null;
fi
done
Este é um exemplo de um arquivo all.rules:
alert udp $HOME_NET any -> any 53 (msg:"ET
TROJAN CopyKittens? Matryoshka DNS Lookup 1 (winupdate64 . com)";
content:"|01 00 00 01 00 00 00 00 00 00|"; depth:10; offset:2;
content:"|0b|winupdate64|03|com|00|"; nocase; distance:0; fast_pattern;
reference:url,www.clearskysec.com/wp-
content/uploads/2017/07/Operation_Wilted_Tulip.pdf; classtype:trojan-
activity; sid:2024495; rev:1;)
alert udp $HOME_NET any -> any 53 (msg:"ET
TROJAN CopyKittens? Matryoshka DNS Lookup 2 (twiter-statics . info)";
content:"|01 00 00 01 00 00 00 00 00 00|"; depth:10; offset:2;
content:"|0e|twiter|2d|statics|04|info|00|"; nocase; distance:0;
fast_pattern; reference:url,www.clearskysec.com/wp-
content/uploads/2017/07/Operation_Wilted_Tulip.pdf;
reference:md5,752240cddda5acb5e8d026cef82e2b54; classtype:trojan-
activity; sid:2024496; rev:1;)
alert udp $HOME_NET any -> any 53 (msg:"ET TROJAN CopyKittens? Cobalt
Strike DNS Lookup (cloudflare-analyse . com)"; content:"|01 00 00 01 00
00 00 00 00 00|"; depth:10; offset:2;
content:"|12|cloudflare|2d|analyse|03|com|00|"; nocase; distance:0;
fast_pattern; threshold:type limit, track by_src, count 1, seconds 60;
reference:url,www.clearskysec.com/wp-
content/uploads/2017/07/Operation_Wilted_Tulip.pdf;
reference:md5,752240cddda5acb5e8d026cef82e2b54; classtype:trojan-
activity; sid:2024497; rev:1;)
Aqui está um exemplo do update.rules:
alert tcp $HOME_NET any -> $EXTERNAL_NET $HTTP_PORTS (msg:"ET TROJAN Revcode
RAT CnC"; flow:established,to_server; content:"POST"; http_method;
content:".php"; http_uri; content:"keyauth="; http_client_body;
fast_pattern; depth:8; content:"&key="; http_client_body; distance:0;
content:"&uid="; http_client_body; distance:0; content:!"Referer|3a|";
http_header; content:"WinHttpRequest"; http_header; metadata:
former_category TROJAN; reference:md5,3f652d9bc17a4be3c0e497ea19848344;
classtype:trojan-activity; sid:2024500; rev:1; metadata:affected_product
Windows_XP_Vista_7_8_10_Server_32_64_Bit, attack_target Client_Endpoint,
deployment Perimeter, signature_severity Major, created_at 2017_07_27,
performance_impact Moderate, updated_at 2017_07_27;)
alert udp $HOME_NET any -> any 53 (msg:"ET TROJAN CopyKittens? Matryoshka
DNS Lookup 1 (winupdate64 . com)"; content:"|01 00 00 01 00 00 00 00 00
00|"; depth:10; offset:2; content:"|0b|winupdate64|03|com|00|"; nocase;
distance:0; fast_pattern; reference:url,www.clearskysec.com/wp-
content/uploads/2017/07/Operation_Wilted_Tulip.pdf; classtype:trojan-
activity; sid:2024495; rev:2;)
alert udp $HOME_NET any -> any 53 (msg:"ET TROJAN CopyKittens Matryoshka DNS
Lookup 2 (twiter-statics . info)"; content:"|01 00 00 01 00 00 00 00 00
00|"; depth:10; offset:2; content:"|0e|twiter|2d|statics|04|info|00|";
nocase; distance:0; fast_pattern; metadata: former_category TROJAN;
reference:url,www.clearskysec.com/wp-
content/uploads/2017/07/Operation_Wilted_Tulip.pdf;
reference:md5,752240cddda5acb5e8d026cef82e2b54; classtype:trojan-activity;
sid:2024496; rev:2; metadata:affected_product
Windows_XP_Vista_7_8_10_Server_32_64_Bit, attack_target Client_Endpoint,
deployment Perimeter, signature_severity Major, created_at 2017_07_25,
malware_family Matryoshka, performance_impact Moderate, updated_at
2017_07_25;)
alert udp $HOME_NET any -> any 53 (msg:"ET TROJAN CopyKittens Cobalt Strike
DNS Lookup (cloudflare-analyse . com)"; content:"|01 00 00 01 00 00 00 00 00
00|"; depth:10; offset:2; content:"|12|cloudflare|2d|analyse|03|com|00|";
nocase; distance:0; fast_pattern; threshold:type limit, track by_src, count
1, seconds 60; metadata: former_category TROJAN;
reference:url,www.clearskysec.com/wp-
content/uploads/2017/07/Operation_Wilted_Tulip.pdf;
reference:md5,752240cddda5acb5e8d026cef82e2b54; classtype:trojan-activity;
sid:2024497; rev:2; metadata:affected_product
Windows_XP_Vista_7_8_10_Server_32_64_Bit, attack_target Client_Endpoint,
deployment Perimeter, signature_severity Major, created_at 2017_07_25,
malware_family CobaltStrike, performance_impact Moderate, updated_at
2017_07_26;)
alert tcp $HOME_NET any -> $EXTERNAL_NET $HTTP_PORTS (msg:"ET TROJAN
Win32/BanloadDownloader.XZY Retrieving Payload"; flow:to_server,established;
content:"GET"; http_method; content:"/sosdoudou_V3/"; http_uri;
fast_pattern; content:"WinHttp.WinHttpRequest"; http_header;
content:!"Accept-"; http_header; content:!"Referer|3a 20|"; http_header;
metadata: former_category TROJAN;
reference:md5,98376de10118892f0773617da137c2be
md5,599ea45f5420f948e0836239eb3ce772; classtype:trojan-activity;
sid:2024499; rev:2; metadata:affected_product
Windows_XP_Vista_7_8_10_Server_32_64_Bit, attack_target Client_Endpoint,
deployment Perimeter, signature_severity Major, created_at 2017_07_26,
malware_family Banload, performance_impact Moderate, updated_at 2017_07_26;)
Observe que há três regras que são as mesmas sid:2024497 , sid:2024496 e sid:2024495 , no entanto, o update.rules contém as revisões de atualização. Eu quero substituir a rev antiga dentro do meu arquivo all.rules com a nova rev da regra em minhas update.rules.
Espero que isso ajude ..
Isso funciona para mim em testes mínimos:
#!/bin/zsh
typeset -A rule sidrev
while read -r line; do
sid=${${line/*sid:/}/;*/}
rev=${${line/*rev:/}/;*/}
if [[ "$rev" -gt "$sidrev[$sid]" ]]; then
sidrev[$sid]="$rev"
rule[$sid]="$line"
fi
done
echo -E ${(F)rule}
O script lê as regras do snort no stdin e no stdout produz a última revisão de tudo que leu.
Fiz algumas suposições para obter algum tipo de resposta. Essas suposições sendo:
Enfim, o suficiente com a blabla, para as partes carnudas
#!/bin/bash
cp "$1" "copy_$1" #backup file we're going to change
#then only extract sid + rev (cut -f cols) + sort
#do that for both file and diff them with RCS format > see output of script for example
#and filter to keep only the sid (prob worth testing without grep to see output)
var=($(diff -n <(egrep -oe "sid:[0-9]{0,11}; rev:[0-9]{0,3}" $1|sort -k2) <(egrep -oe "sid:[0-9]{0,11}; rev:[0-9]{0,3}" $2|sort -k 2) | egrep -oe "sid:[0-9]{0,11};" ))
#Now loop over each SID
for i in ${var[@]}; do
#Extract line number in IF1
oldline=$(grep -n $i $1|cut -f1 -d:)
#Extract replacement line in IF2
newline=$(grep $i $2.txt)
#awk magic see (note sed was a pain, couldn't get it to work :/) >> https://sobrelinux.info/questions/4595/how-to-replace-a-string-on-the-5th-line-of-multiple-text-files"$oldline" -v repl="$newline" '(NR==nline){$0=repl}1;' "copy_$1" > "f.tmp" && mv "f.tmp" "copy_$1"
done
#then highlight diff between copy and original // you'll still need to mv "copy_$1" "$1" for it to be applied
diff -n "copy_$1" "$1"
Envolva tudo isso em um script e chame-o:
./my_script.sh all_rules.rules /home/path/update_rules.rules
Você não mencionou perl , mas suponho que não há problema em usar desde que você mencionou awk , sed e grep . Eu escolhi perl em vez de awk principalmente porque é muito mais fácil trabalhar com matrizes associativas de vários níveis (um "Hash-of-Hashes" ou "HoH" em perl lingo) do que em awk .
#!/usr/bin/perl
use strict;
# array used to keep track of the order each sid was first seen,
# so that they can be printed out in the same order.
# Necessary because perl hashes are inherently un-ordered.
my @order=();
# hashed array to contain the highest rev seen of each sid.
my %S = ();
# count of the number of files we've read completely so far.
my $filenum=0;
while(<>) {
s/^\s*|\s*$//g; # strip leading and trailing spaces
if (m/^$/) { $filenum++ if eof; next }; # skip empty lines
# extract the sid and the rev
my ($sid, $rev) = $_ =~ (m/^.*; sid:(\d+); rev:(\d+)/) ;
# store or update an anonymous hash containing the rev and the entire
# line in the hash, keyed by the sid.
if (defined($S{$sid})) {
$S{$sid} = { rev => $rev, line => $_ } if ( ($rev > $S{$sid}->{rev}) );
} else {
next if ($filenum); # only store sid if we're still reading the 1st file.
push @order, $sid;
$S{$sid} = { rev => $rev, line => $_ };
};
$filenum++ if eof;
};
# if you want output sorted by the sid, comment the first of the next
# two lines and uncomment the second
for my $sid (@order) {
#for my $sid (sort keys %S) {
print $S{$sid}->{line}, "\n";
};
Salve-o como, por exemplo, ./apply-update.pl e execute-o como ./apply-update.pl all.rules update.rules > out.rules
A saída dos novos arquivos de entrada de amostra será de apenas três linhas (com os sids vistos em all.rules , atualizados por update.rules , ou seja, as versões "rev: 2" em vez de "rev: 1"):
alert udp $home_net any -> any 53 (msg:"et trojan copykittens? matryoshka dns lookup 1 (winupdate64 . com)"; content:"|01 00 00 01 00 00 00 00 00 00|"; depth:10; offset:2; content:"|0b|winupdate64|03|com|00|"; nocase; distance:0; fast_pattern; reference:url,www.clearskysec.com/wp- content/uploads/2017/07/operation_wilted_tulip.pdf; classtype:trojan- activity; sid:2024495; rev:2;)
alert udp $home_net any -> any 53 (msg:"et trojan copykittens matryoshka dns lookup 2 (twiter-statics . info)"; content:"|01 00 00 01 00 00 00 00 00 00|"; depth:10; offset:2; content:"|0e|twiter|2d|statics|04|info|00|"; nocase; distance:0; fast_pattern; metadata: former_category trojan; reference:url,www.clearskysec.com/wp- content/uploads/2017/07/operation_wilted_tulip.pdf; reference:md5,752240cddda5acb5e8d026cef82e2b54; classtype:trojan-activity; sid:2024496; rev:2; metadata:affected_product windows_xp_vista_7_8_10_server_32_64_bit, attack_target client_endpoint, deployment perimeter, signature_severity major, created_at 2017_07_25, malware_family matryoshka, performance_impact moderate, updated_at 2017_07_25;)
alert udp $home_net any -> any 53 (msg:"et trojan copykittens cobalt strike dns lookup (cloudflare-analyse . com)"; content:"|01 00 00 01 00 00 00 00 00 00|"; depth:10; offset:2; content:"|12|cloudflare|2d|analyse|03|com|00|"; nocase; distance:0; fast_pattern; threshold:type limit, track by_src, count 1, seconds 60; metadata: former_category trojan; reference:url,www.clearskysec.com/wp- content/uploads/2017/07/operation_wilted_tulip.pdf; reference:md5,752240cddda5acb5e8d026cef82e2b54; classtype:trojan-activity; sid:2024497; rev:2; metadata:affected_product windows_xp_vista_7_8_10_server_32_64_bit, attack_target client_endpoint, deployment perimeter, signature_severity major, created_at 2017_07_25, malware_family cobaltstrike, performance_impact moderate, updated_at 2017_07_26;)
Isso será muitas vezes mais rápido do que qualquer versão bash - os interpretadores de shell não estão nem um pouco perto do processamento de texto como perl ou awk ou python . Uma linguagem compilada como C seria muito mais rápida novamente, mas escrever algo assim em C seria provavelmente pelo menos 50 ou 60, talvez até algumas centenas, linhas de código C, não 15 linhas de perl (sem contar os comentários). ou linhas em branco).