From lists at keamera.org Tue Sep 1 04:52:34 2009 From: lists at keamera.org (Guido Landi) Date: Tue, 01 Sep 2009 10:52:34 +0200 Subject: [Emerging-Sigs] IIS-FTP exploit signature (milw0rm) released In-Reply-To: <20090831235136.GA40797@knobbe.us> References: <1251755569.42741.11.camel@localhost> <4A9C5FF2.9090009@packetmail.net> <4A9C6159.4050903@packetmail.net> <20090831235136.GA40797@knobbe.us> Message-ID: <4A9CE0D2.1040308@keamera.org> Hello Frank, actually the SITE command is not related to the vulnerability and it is used in this exploit as a placeholder for the shellcode. This means that also your signature is exploit specific and variants of this exploit could no be detected this way. A more effective way to detect a generic exploit attempt would be to check for large NLST commands with '*'. Guido. Frank Knobbe wrote: > On Mon, Aug 31, 2009 at 06:48:41PM -0500, evilghost at packetmail.net wrote: >> Ugh. Please throw out the first sig I wrote the 'VVVV' after the KSEXY, >> this was an error on my part. Concatenated the content matches, threw >> out nocase since the Perl code prints all uppercase "SITE" inclusive of >> the KSEXY padding. I don't imagine this sig will false much, not likely >> we'll see "SITE KSEXY" often on FTPd. >> >> alert tcp $EXTERNAL_NET any -> $HOME_NET 21 (msg:"ET Exploit IIS FTP Exploit - Large SITE Command, milw0rm PoC"; flow:established,to_server; content:"SITE KSEXY"; classtype:attempted-admin; reference:url,www.milw0rm.com/exploits/9541; sid:2009xxxx; rev:0;) > > Yeah, but I'm not a big fan of exploit specific signatures. It's easy to > change the KSEXY to KDULL to avoid detection :) > > -Frank > > _______________________________________________ > Emerging-sigs mailing list > Emerging-sigs at emergingthreats.net > http://lists.emergingthreats.net/mailman/listinfo/emerging-sigs > From mike.lococo at nyu.edu Tue Sep 1 00:39:33 2009 From: mike.lococo at nyu.edu (Mike Lococo) Date: Tue, 01 Sep 2009 00:39:33 -0400 Subject: [Emerging-Sigs] HTTP Exe Download Sig In-Reply-To: <4A9C7E0A.4050800@jonkmans.com> References: <4A9C2A5C.7080201@gmail.com> <4A9C7E0A.4050800@jonkmans.com> Message-ID: <4A9CA585.5020303@nyu.edu> > We have a few variations of the idea, it's definitely a good one when > used where load allows: > > Take a look at: > www.emergingthreats.net/cgi-bin/cvsweb.cgi/sigs/POLICY/POLICY_EXE > www.emergingthreats.net/cgi-bin/cvsweb.cgi/sigs/POLICY/POLICY_EXE_HTTP > www.emergingthreats.net/cgi-bin/cvsweb.cgi/sigs/POLICY/POLICY_EXE_NoUserAgent > www.emergingthreats.net/cgi-bin/cvsweb.cgi/sigs/POLICY/POLICY_EXE_in_BMP Thanks all for the tips on alternate (and better) approaches. And thanks Matt for pointing out similar existing rules. I have historically steered clear of the policy rules, which are disproportionately likely to make me stumble on content that makes my eyes bleed. I'll likely replace my local rule with some combination of the above options, though. > Do be careful running these in a high load environment. But I definitely > agree they're valuable. Entirely agreed that they can impact a heavily loaded sensor. If you have some headroom, I find it a worthwhile use of resource though. Thanks, Mike Lococo From jonkman at jonkmans.com Tue Sep 1 08:47:00 2009 From: jonkman at jonkmans.com (Matt Jonkman) Date: Tue, 01 Sep 2009 08:47:00 -0400 Subject: [Emerging-Sigs] some new rules for an old exploit In-Reply-To: <20090831160517.8zpsfj4su80ogk08@mail.afferentsecurity.com> References: <20090831160517.8zpsfj4su80ogk08@mail.afferentsecurity.com> Message-ID: <4A9D17C4.80604@jonkmans.com> Hey Jack, questions inline: Jack Pepper wrote: > Looking at a compromised server found these items: > > alert tcp $HOME_NET any -> $EXTERNAL_NET any (msg:"Local server has > been rooted with zetha web shell"; flow:from_server,established; > content:"|3c|title|3e|ZETHA WEB SHELL"; classtype: successful-admin; > sid:1082331; rev:1;) Safe to go http_servers with the source side? Can we limit to web ports, or is this one injected and used on off ports? > > alert tcp $HOME_NET any -> $EXTERNAL_NET $HTTP_PORTS (msg:"Local > server compromised downloading injectable rootkit"; > flow:from_server,established; > content:"Loader'z"; content:"Pro-Hack.ru"; nocase;classtype: > successful-admin; sid:1082332; rev:1;) > Posting > The exploit that was used to get in was a php inclusion in Apache > appserv from 2006. Ref: http://securityvulns.com/Kdocument914.html . > Logs show repeated ocurrences of the leading double slash > > alert tcp $EXTERNAL_NET any -> $HOME_NET $HTTP_PORTS (msg:"ET WEB > Apache Appserv > Injection Attempt"; flow:established,to_server; > uricontent: "//appserv/main.php"; nocase; classtype:attempted-admin; > ref:url,securityvulns.com/Kdocument914.html; sid:1082333; rev:1;) > The doubel quotes will be normalized out by the preproc. That's a pretty slim thing to alert on though. Anything else we can look for here? Matt -- -------------------------------------------- Matthew Jonkman Emerging Threats Open Information Security Foundation (OISF) Phone 765-429-0398 Fax 312-264-0205 http://www.emergingthreats.net http://www.openinformationsecurityfoundation.org -------------------------------------------- PGP: http://www.jonkmans.com/mattjonkman.asc From jonkman at jonkmans.com Tue Sep 1 08:54:00 2009 From: jonkman at jonkmans.com (Matt Jonkman) Date: Tue, 01 Sep 2009 08:54:00 -0400 Subject: [Emerging-Sigs] some new rules for an old exploit In-Reply-To: <20090831160517.8zpsfj4su80ogk08@mail.afferentsecurity.com> References: <20090831160517.8zpsfj4su80ogk08@mail.afferentsecurity.com> Message-ID: <4A9D1968.5030901@jonkmans.com> Jack Pepper wrote: > > alert tcp $HOME_NET any -> $EXTERNAL_NET $HTTP_PORTS (msg:"Local > server compromised downloading injectable rootkit"; > flow:from_server,established; > content:"Loader'z"; content:"Pro-Hack.ru"; nocase;classtype: > successful-admin; sid:1082332; rev:1;) Actually on this, can we do a depth and within? What's the phrase look like? Matt > > The exploit that was used to get in was a php inclusion in Apache > appserv from 2006. Ref: http://securityvulns.com/Kdocument914.html . > Logs show repeated ocurrences of the leading double slash > > alert tcp $EXTERNAL_NET any -> $HOME_NET $HTTP_PORTS (msg:"ET WEB > Apache Appserv > Injection Attempt"; flow:established,to_server; > uricontent: "//appserv/main.php"; nocase; classtype:attempted-admin; > ref:url,securityvulns.com/Kdocument914.html; sid:1082333; rev:1;) > > jp > -- -------------------------------------------- Matthew Jonkman Emerging Threats Open Information Security Foundation (OISF) Phone 765-429-0398 Fax 312-264-0205 http://www.emergingthreats.net http://www.openinformationsecurityfoundation.org -------------------------------------------- PGP: http://www.jonkmans.com/mattjonkman.asc From evilghost at packetmail.net Tue Sep 1 08:59:48 2009 From: evilghost at packetmail.net (evilghost@packetmail.net) Date: Tue, 1 Sep 2009 07:59:48 -0500 Subject: [Emerging-Sigs] IIS-FTP exploit signature (milw0rm) released In-Reply-To: <4A9CE0D2.1040308@keamera.org> References: <1251755569.42741.11.camel@localhost> <4A9C5FF2.9090009@packetmail.net> <4A9C6159.4050903@packetmail.net> <20090831235136.GA40797@knobbe.us> <4A9CE0D2.1040308@keamera.org> Message-ID: <4A9D1AC4.40402@packetmail.net> What is everyones thought on this signature? Looking specifically at the milw0rm exploit I see the data passed to NLST as 178 characters long, setting the within:100 seems reasonable. Perhaps a |2a 2f| instead of just |2a|? alert tcp $EXTERNAL_NET any -> $HOME_NET 21 (msg:"ET Exploit IIS FTP Exploit - NLST"; flow:established,to_server; content:"NLST"; nocase; content:"|2a|"; content:!"|0d 0a|"; within:100; classtype:attempted-admin; reference:url,www.milw0rm.com/exploits/9541; sid:2009xxxx; rev:0;) -evilghost Guido Landi wrote: > Hello Frank, > > actually the SITE command is not related to the vulnerability and it is > used in this exploit as a placeholder for the shellcode. This means that > also your signature is exploit specific and variants of this exploit > could no be detected this way. > > A more effective way to detect a generic exploit attempt would be to > check for large NLST commands with '*'. > > > Guido. > > Frank Knobbe wrote: > >> On Mon, Aug 31, 2009 at 06:48:41PM -0500, evilghost at packetmail.net wrote: >> >>> Ugh. Please throw out the first sig I wrote the 'VVVV' after the KSEXY, >>> this was an error on my part. Concatenated the content matches, threw >>> out nocase since the Perl code prints all uppercase "SITE" inclusive of >>> the KSEXY padding. I don't imagine this sig will false much, not likely >>> we'll see "SITE KSEXY" often on FTPd. >>> >>> alert tcp $EXTERNAL_NET any -> $HOME_NET 21 (msg:"ET Exploit IIS FTP Exploit - Large SITE Command, milw0rm PoC"; flow:established,to_server; content:"SITE KSEXY"; classtype:attempted-admin; reference:url,www.milw0rm.com/exploits/9541; sid:2009xxxx; rev:0;) >>> >> Yeah, but I'm not a big fan of exploit specific signatures. It's easy to >> change the KSEXY to KDULL to avoid detection :) >> >> -Frank >> >> _______________________________________________ >> Emerging-sigs mailing list >> Emerging-sigs at emergingthreats.net >> http://lists.emergingthreats.net/mailman/listinfo/emerging-sigs >> >> > > _______________________________________________ > Emerging-sigs mailing list > Emerging-sigs at emergingthreats.net > http://lists.emergingthreats.net/mailman/listinfo/emerging-sigs > > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.emergingthreats.net/pipermail/emerging-sigs/attachments/20090901/6ffbb0c1/attachment.html From eslerj at gmail.com Tue Sep 1 09:16:30 2009 From: eslerj at gmail.com (Joel Esler) Date: Tue, 1 Sep 2009 09:16:30 -0400 Subject: [Emerging-Sigs] IIS-FTP exploit signature (milw0rm) released In-Reply-To: <4A9D1AC4.40402@packetmail.net> References: <1251755569.42741.11.camel@localhost> <4A9C5FF2.9090009@packetmail.net> <4A9C6159.4050903@packetmail.net> <20090831235136.GA40797@knobbe.us> <4A9CE0D2.1040308@keamera.org> <4A9D1AC4.40402@packetmail.net> Message-ID: <314cf0830909010616n92795f1n9f1e299c47b17baa@mail.gmail.com> Actually, (just off the top of my head, I haven't checked AT ALL) if your default def_max_param_len in the ftp_telnet preprocessor is still set to 100, won't the ftp_telnet preprocessor catch this by default? NLST is in the def_max_param_len for 100. Check README.ftptelnet. J On Tue, Sep 1, 2009 at 8:59 AM, evilghost at packetmail.net < evilghost at packetmail.net> wrote: > What is everyones thought on this signature? Looking specifically at the > milw0rm exploit I see the data passed to NLST as 178 characters long, > setting the within:100 seems reasonable. Perhaps a |2a 2f| instead of just > |2a|? > > alert tcp $EXTERNAL_NET any -> $HOME_NET 21 (msg:"ET Exploit IIS FTP Exploit - NLST"; flow:established,to_server; content:"NLST"; nocase; content:"|2a|"; content:!"|0d 0a|"; within:100; classtype:attempted-admin; reference:url,www.milw0rm.com/exploits/9541; sid:2009xxxx; rev:0;) > > -evilghost > > > Guido Landi wrote: > > Hello Frank, > > actually the SITE command is not related to the vulnerability and it is > used in this exploit as a placeholder for the shellcode. This means that > also your signature is exploit specific and variants of this exploit > could no be detected this way. > > A more effective way to detect a generic exploit attempt would be to > check for large NLST commands with '*'. > > > Guido. > > Frank Knobbe wrote: > > > On Mon, Aug 31, 2009 at 06:48:41PM -0500, evilghost at packetmail.net wrote: > > > Ugh. Please throw out the first sig I wrote the 'VVVV' after the KSEXY, > this was an error on my part. Concatenated the content matches, threw > out nocase since the Perl code prints all uppercase "SITE" inclusive of > the KSEXY padding. I don't imagine this sig will false much, not likely > we'll see "SITE KSEXY" often on FTPd. > > alert tcp $EXTERNAL_NET any -> $HOME_NET 21 (msg:"ET Exploit IIS FTP Exploit - Large SITE Command, milw0rm PoC"; flow:established,to_server; content:"SITE KSEXY"; classtype:attempted-admin; reference:url,www.milw0rm.com/exploits/9541; sid:2009xxxx; rev:0;) > > > Yeah, but I'm not a big fan of exploit specific signatures. It's easy to > change the KSEXY to KDULL to avoid detection :) > > -Frank > > _______________________________________________ > Emerging-sigs mailing listEmerging-sigs at emergingthreats.nethttp://lists.emergingthreats.net/mailman/listinfo/emerging-sigs > > _______________________________________________ > Emerging-sigs mailing listEmerging-sigs at emergingthreats.nethttp://lists.emergingthreats.net/mailman/listinfo/emerging-sigs > > > _______________________________________________ > Emerging-sigs mailing list > Emerging-sigs at emergingthreats.net > http://lists.emergingthreats.net/mailman/listinfo/emerging-sigs > > -------------- next part -------------- An HTML attachment was scrubbed... URL: http://lists.emergingthreats.net/pipermail/emerging-sigs/attachments/20090901/81d6520e/attachment.html From jonkman at jonkmans.com Tue Sep 1 09:22:58 2009 From: jonkman at jonkmans.com (Matt Jonkman) Date: Tue, 01 Sep 2009 09:22:58 -0400 Subject: [Emerging-Sigs] 2 more sigs In-Reply-To: References: <4A9BDF2B.9000603@jonkmans.com> Message-ID: <4A9D2032.2060001@jonkmans.com> Unfortunately I don't think that helps any. :) Just higher load now. I think we'll have to pass on this concept for now... Kevin Ross wrote: >>> Forgot to send this to the emerging-sigs list also for comment :) > > > What about this then? > > alert tcp $EXTERNAL_NET any -> $HTTP_SERVERS $HTTP_PORTS (msg:"WEB > Cross Site Scripting Attempt in Cookie (