More Snort News

I have several developments to report from the Snort front. First, Jeremy Hewlett announced Thursday the release of Snort 2.3.2. This version is a quick response to the problem parsing Bleeding Snort rules reported shortly after Snort 2.3.1 arrived. I think this release was quickly pushed out the door to demonstrate that Sourcefire was not trying to lock out Bleeding Snort users. This is smart; there's no need to repeat a Microsoft-style "DOS isn't done until Lotus won't run" situation with Snort!

Speaking of Bleeding Snort, Matt Jonkman announced Friday work on a new "Open Source Snort Rules Consortium." He says:

"The OSSRC will be a group that any company or organization will be welcome to join. The members will share research on new threats and rules to handle those threats, with the goal of creating a unified community-based ruleset. Each member may post these rules wherever they choose, distribute them to their clients or customers, or use them in their own subscription services according to the provisions in the GPL. The goals of the group are still forming, but initially will be to:

1. Maintain a fast moving and GPL-licensed Snort ruleset
2. Avoid rule duplication amongst community rulesets, both in terms of content and SIDs
3. Improve and enforce quality standards for rules (documentation, etc.)
4. Possibly move to a Stable and Unstable rule 'vetting' process"

It will be interesting to see how this rule set progresses. If you visit the Snort rules download page you'll see three sets of rules:

- Sourcefire VRT Certified Rules - The Official Snort Ruleset (subscription release)

- Sourcefire VRT Certified Rules - The Official Snort Ruleset

- Community Rules

The first are the latest and greatest, available to subscribers. The second are the five-day-delay version of the first set, available to registered users. The third are new, and include:

- community-exploit.rules
- community-ftp.rules
- community-game.rules
- community-inappropriate.rules
- community-mail-client.rules
- community-sql-injection.rules
- community-virus.rules
- community-web-cgi.rules
- community-web-client.rules
- community-web-dos.rules
- community-web-misc.rules

A look at the sid-msg.map summarizes the rules included in this first community rule set. Observe that some of the files (community-web-dos.rules, for example) are empty:

# Copyright 2005 Sourcefire, Inc. All Rights Reserved.
# This file is licensed under the GNU General Public License.
# Please see the file LICENSE in this directory for more details.
# Id SID -> MSG map

100000100 || COMMUNITY EXPLOIT Windows Acrobat Reader Activex Overflow Flowbit || cve,2004-0629 || bugtraq,10947
100000101 || COMMUNITY EXPLOIT Windows Acrobat Reader Activex Overflow Exploit || cve,2004-0629 || bugtraq,10947
100000102 || COMMUNITY GAME Halocon Denial of Service Empty UDP Packet || bugtraq,12281
100000103 || COMMUNITY GAME Breed Game Server Denial of Service Empty UDP Packet || bugtraq,12262
100000104 || COMMUNITY GAME Amp II 3D Game Server Denial of Service Empty UDP Packet || bugtraq,12192
100000105 || INAPPROPRIATE lolita sex
100000106 || COMMUNITY SQL-INJECTION Microsoft BizTalk Server 2002 rawdocdata.asp || bugtraq,7470 || cve,2003-0118 || url,www.microsoft.com/technet/security/bulletin/MS03-016.mspx
100000107 || COMMUNITY SQL-INJECTION Microsoft BizTalk Server 2002 RawCustomSearchField.asp || bugtraq,7470 || cve,2003-0118 || url,www.microsoft.com/technet/security/bulletin/MS03-016.mspx
100000108 || COMMUNITY SQL-INJECTION OpenBB board.php || bugtraq,7404
100000109 || COMMUNITY SQL-INJECTION OpenBB member.php || bugtraq,7404
100000110 || COMMUNITY VIRUS Dabber PORT overflow attempt port 5554 || MCAFEE,125300
100000111 || COMMUNITY VIRUS Dabber PORT overflow attempt port 1023 || MCAFEE,125300
100000112 || WEB-CGI Readfile.tcl Access || bugtraq,7426
100000113 || COMMUNITY WEB-CGI HappyMall Command Execution member_html.cgi || bugtraq,7530 || cve,2003-0243
100000114 || COMMUNITY WEB-CGI HappyMall Command Execution normal_html.cgi || bugtraq,7530 || cve,2003-0243
100000115 || COMMUNITY WEB-CGI PHP-Nuke Web_Links Path Disclosure Null CID || bugtraq,7589
100000116 || COMMUNITY WEB-CGI PHP-Nuke Web_Links Path Disclosure Non-Numeric CID || bugtraq,7589
100000117 || COMMUNITY WEB-CGI VBulliten Remote Command Execution Attempt || bugtraq,12542
100000118 || WEB-CLIENT Internet Explorer URLMON.DLL Content-Type Overflow Attempt || bugtraq,7419 || cve,2003-0113 || url,www.microsoft.com/technet/security/bulletin/MS03-015.mspx
100000119 || WEB-CLIENT Internet Explorer URLMON.DLL Content-Encoding Overflow Attempt || bugtraq,7419 || cve,2003-0113 || url,www.microsoft.com/technet/security/bulletin/MS03-015.mspx
100000121 || COMMUNITY WEB-MISC Test Script Access
100000122 || COMMUNITY WEB-MISC mod_jrun overflow attempt || bugtraq,11245 || cve,2004-0646
100000123 || INAPPROPRIATE preteen sex
100000124 || INAPPROPRIATE girls gone wild

That's only 24 rules at the moment. Let's look at the first two:

alert tcp $HOME_NET any -> $EXTERNAL_NET $HTTP_PORTS
(msg:"COMMUNITY EXPLOIT Windows Acrobat Reader Activex Overflow Flowbit";
flow:to_server,established; pcre:"/.{1050,}/U"; flowbits:set,uri.size.1050;
flowbits:noalert; reference:cve,2004-0629; reference: bugtraq,10947;
classtype:attempted-user; sid: 100000100; rev:1;)

alert tcp $EXTERNAL_NET $HTTP_PORTS -> $HOME_NET any
(msg:"COMMUNITY EXPLOIT Windows Acrobat Reader Activex Overflow Exploit";
flow:to_client,established; content:"Content-Type|3A|"; nocase;
pcre:"/^Content-Type\x3a\s*application\x2f(pdf|vnd\x2efdf|
vnd\x2eadobe\x2exfdf|vnd\x2eadobe\x2exdp+xml|vnd\x2e\ adobe\x2exfd+xml)/smi";
flowbits:isset,uri.size.1050; reference:cve,2004-0629; reference:bugtraq,10947;
classtype:attempted-user; sid:100000101; rev:1;)


When I met with Marty on Thursday, he said that rules like the first one that invoke PCRE but do not use a content match really slow down the detection engine. The second rule also uses (complicated) PCRE, but there is a content match.

Given the limited number of rules in this community set, I can see why a company like StillSecure decided to sponsor Bleeding Snort. StillSecure's IDS is built on Snort, so they have an incentive to sponsor signature research.

Comments

Anonymous said…
This comment has been removed by a blog administrator.

Popular posts from this blog

Zeek in Action Videos

New Book! The Best of TaoSecurity Blog, Volume 4

MITRE ATT&CK Tactics Are Not Tactics