I’m going to be using NSM-Console version 0.4-DEVEL, which adds the features that allow this analysis to be performed without external tools. You can get the development version here. Alright, let’s get this party started:

First things first, the fellows at SANS point you to the first packet in the xmas_Starter.pcap file, so let’s load up NSM-Console with the packet capture

./nsm ~/xmas_Starter.pcap

Next, let’s do a printout of all the packets in this dump (since it’s a small file, there shouldn’t be too many)

nsm> p *
Args: *
Filename: /Users/hinmanm/xmas_Starter.pcap
list from 1 to *
1 1198471642.61773 192.168.25.1 -> 192.168.25.255 UDP 138 > 138 Len=243
2 1198471662.79806 192.168.25.100 -> 192.168.25.128 TCP 7337 > 1000 Len=254
3 1198471662.79813 192.168.25.128 -> 192.168.25.100 TCP 1000 > 7337 Len=58
4 1198471662.79877 192.168.25.100 -> 192.168.25.128 TCP 7337 > 1000 Len=60
5 1198471663.79691 192.168.25.100 -> 192.168.25.128 TCP 7337 > 1000 Len=254
6 1198471663.79697 192.168.25.128 -> 192.168.25.100 TCP 1000 > 7337 Len=58
… etc, etc, etc, to a total of 25 packets

Well, the SANS guide says to start at packet #1, so let’s take a look at it

nsm> p -x 1
(see image for output)

Whoops, this doesn’t look like useful output, maybe they meant the first TCP packet? (Instead of UDP). Looks like the first TCP packet is #2, so let’s look at that one:

nsm> p -x 2
Args: 2
Filename: /Users/hinmanm/xmas_Starter.pcap
full from 2 to 2
2 1198471662.79806 192.168.25.100 -> 192.168.25.128 TCP 7337 > 1000 Len=254
0010 53 57 34 67 64 47 68 6c 49 47 31 76 64 6d 6c 6c SW4gdGhlIG1vdmll
0020 49 45 45 67 51 32 68 79 61 58 4e 30 62 57 46 7a IEEgQ2hyaXN0bWFz
0030 49 45 4e 68 63 6d 39 73 4c 43 42 6f 62 33 63 67 IENhcm9sLCBob3cg
0040 62 57 46 75 65 53 42 75 61 57 64 6f 64 43 68 7a bWFueSBuaWdodChz
0050 4b 53 42 6b 61 57 51 67 64 47 68 6c 49 48 52 6f KSBkaWQgdGhlIHRo
0060 63 6d 56 6c 49 48 4e 77 61 58 4a 70 64 48 4d 67 cmVlIHNwaXJpdHMg
0070 59 32 39 74 5a 53 42 30 62 79 42 32 61 58 4e 70 Y29tZSB0byB2aXNp
0080 64 44 38 3d 00 00 00 00 00 00 00 00 00 00 00 00 dD8=............
0090 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00a0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00b0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................
00c0 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 ................

Ahh, that looks like more readable output, due to the string ending with “=”, it’s most likely encoded in base64 (since = is used for padding). Let’s output the string in just plain ascii so we can decode it easier

nsm> p -a 2
Args: 2
Filename: /Users/hinmanm/xmas_Starter.pcap
ascii from 2 to 2
2 1198471662.79806 192.168.25.100 -> 192.168.25.128 TCP 7337 > 1000 Len=254
SW4gdGhlIG1vdmllIEEgQ2hyaXN0bWFzIENhcm9sLCBob3cgbWFueSBuaWdodChz KSBkaWQgdGhlIHRocmVlIHNwaXJpdHMgY29tZSB0byB2aXNpdD8=...................... ..............................................................

(I put line breaks in so it wouldn’t distort the page)
Okay, let’s take this line and decode it to get our first clue

nsm> decode base64 SW4gdGhlIG1vdmllIEEgQ2hyaXN0bWFzIENhcm9sLCBob3cgbWFueSBuaWdodChz KSBkaWQgdGhlIHRocmVlIHNwaXJpdHMgY29tZSB0byB2aXNpdD8=
Decoding base64 --> ascii...
Output ([]'s added to show beginning and end):

[In the movie A Christmas Carol, how many night(s) did the three spirits come to visit?]

Alright! Got the clue! Now let’s load up our answer file and get the next one

nsm> file /Users/hinmanm/xmas_challenge_2007.pcap
Setting ${PCAP_FILE} = /Users/hinmanm/xmas_challenge_2007.pcap
Setting ${PCAP_BASE} = xmas_challenge_2007.pcap

Since the 3 spirits came on 1 night, we know our next answer is in packet #1, so let’s take a look

nsm> p -a 1
Args: 1
Filename: /Users/hinmanm/xmas_challenge_2007.pcap
ascii from 1 to 1
1 1194153111.12232 192.168.25.100 -> 192.168.25.128 TCP 7337 > 1000 Len=154
QWxsIEkgd2FudCBmb3IgQ2hyaXN0bWFzIGlzIG15IF9fX18gRnJvbnQgVGVldG gu....................................

Aha! Another base64 encoding, let’s decode it

nsm> decode base64 QWxsIEkgd2FudCBmb3IgQ2hyaXN0bWFzIGlzIG15 IF9fX18gRnJvbnQgVGVldGgu
Decoding base64 --> ascii...
Output ([]'s added to show beginning and end):

[All I want for Christmas is my ____ Front Teeth.]

Alright, I think this shows the basic idea, and since this post is to illustrate some of the new features of NSM-Console (rather than the solution to the puzzle), let’s skip on ahead.

Packet 2 leads you to packet 3
Packet 3 leads you to packet 9
Packet 9 leads you to packet 11
Packet 11 leads you to packet 12
Packet 12 leads you to packet 359

Here’s where things start to get a little more interesting, printing out packet 359 shows that it doesn’t look like it’s encoded base64 anymore, in fact it looks like it’s urlescaped

nsm> p -x 359
Args: 359
Filename: /Users/hinmanm/xmas_challenge_2007.pcap
full from 359 to 359
359 1194153771.83615 192.168.25.100 -> 192.168.25.128 TCP 7337 > 1000 Len=154
0010 38 37 25 32 30 31 30 31 25 32 30 4e 55 4c 4c 25 87%20101%20NULL%
0020 32 30 31 31 39 25 32 30 31 30 35 25 32 30 31 31 20119%20105%2011
0030 35 25 32 30 31 30 34 25 32 30 4e 55 4c 4c 25 32 5%20104%20NULL%2
0040 30 31 32 31 25 32 30 31 31 31 25 32 30 31 31 37 0121%20111%20117
0050 25 32 30 4e 55 4c 4c 25 32 30 39 37 25 32 30 4e %20NULL%2097%20N
0060 55 4c 4c 25 32 30 37 37 25 32 30 31 30 31 25 32 ULL%2077%20101%2

Also, the SANS challenge mentions that the message may or may not be in multiple packets, so let’s check the next one

nsm> p -x 360
Args: 360
Filename: /Users/hinmanm/xmas_challenge_2007.pcap
full from 360 to 360
360 1194153772.83062 192.168.25.100 -> 192.168.25.128 TCP 7337 > 1000 Len=154
0010 25 32 30 31 31 34 25 32 30 31 32 31 25 32 30 4e %20114%20121%20N
0020 55 4c 4c 25 32 30 36 37 25 32 30 31 30 34 25 32 ULL%2067%20104%2
0030 30 31 31 34 25 32 30 31 30 35 25 32 30 31 31 35 0114%20105%20115
0040 25 32 30 31 31 36 25 32 30 31 30 39 25 32 30 39 %20116%20109%209
0050 37 25 32 30 31 31 35 25 32 30 34 34 25 30 44 25 7%20115%2044%0D%
0060 30 41 38 37 25 32 30 31 30 31 25 32 30 4e 55 4c 0A87%20101%20NUL

Looks like a continuation, if you print out a few more, you find out that the data stops in packet #365, let’s print out all the packets so we can see what the data looks like

nsm> p -x 359-365
(see picture for output, too long to paste here)

Alright, looks like we’ve got our message, let’s decode it

nsm> p -a 359-365
gives us:

nsm> decode urlescape <big long escaped text>
shows:

That looks like decimal ascii values, let’s decode the values using the “char” decoding

nsm> decode char <space separated char codes>
Here’s the output from decoding each of the 3 lines:

And look! There’s our message (I replaced all the ‘NULL’s with spaces):

We wish you a Merry Christmas,
We wish you a Merry Christmas,
We wish you a Merry Christmas
and a Happy New Year!!!

Merry Christmas to you too SANS, and thanks for the awesome challenge!

You can check out SANS’ solution for the challenge here.
You can check out Geek00l’s full solution for the challenge here.

If it wasn’t for this challenge, we might not have had the idea to include these features in NSM-Console, I’m glad we did and I hope it proves useful to the rest of the packet monkeys out there

P.S. A big thank-you to Scholar for letting me use his pcap parsing library, thus eliminating the dependency on any 3rd party libraries to do all the packet reading, thanks a bunch!

Have another idea for a feature that should be in NSM-Console? Shoot me an email or leave a comment!

#!/usr/bin/env ruby
def fisher_yates_shuffle(a)
(a.size-1).downto(1) { |i|
j = rand(i+1)
a[i], a[j] = a[j], a[i] if i != j
}
end
lines = File.open('/usr/share/dict/words').collect
fisher_yates_shuffle(lines)
lines.each {
|word|
puts "trying #{word.chomp}..."
system("wget -q #{ARGV[0]}/#{word.chomp}.html")
system("wget -q #{ARGV[0]}/#{word.chomp}.htm")
system("wget -q #{ARGV[0]}/#{word.chomp}.php")
sleep(1)
}

(The “sleep(1)” is so you don’t kill the server with traffic, remove if you like)

Should be pretty self-explainatory, go through all the words in /usr/share/dict/words and attempt to fetch webpages. At my current word dictionary size, it would take 65 hours to complete (1 second per word)

A smart person would replace “/usr/share/dict/words” in the script with a better list of website pagenames, if they actually wanted to use this

You know, I’ve always wondered if servers had a “rurigenous.html” or a “mastochondroma.php” webpage on their site…