ddaa's blog
Write-ups for CTF.
ddaa's blog

Trend Micro CTF 2017 write-ups

engima , wireshark , frequency analysis , Other CTF

Share Tweet Share

Our team phddaa (what the f...) got 2400 points and 19th rank this year.
We didn't spend too much time on this game because we think the competition is 48 hours. O__O
There are several categories of challenges but I don't know how they distinguish.
Almost of challenges need to analyze and guess... Thus, I put this write up in misc category.

Analysis-Offensive 100

The problem provided a binary named Forensic_Encyption. The file type is MS-DOS but it's not a real MS-DOS executable. After a liitle guessing, I found the binary is a zip file. We can get two files, file_1 and file_2 after extracted Forensic_Encyption.

  • file_1
    • An image with jpeg format hide a string VHVyaW5nX01hY2hpbmVfYXV0b21hdG9u in exif information.
    • Decode the string with base64 and get Turing_Machine_automaton.
  • file_2
    • Another zip file.
    • We can extract a text file key.txt with the password Turing_Machine_automaton.

key.txt is a file which recorded the information about ipsec. I spent some time at this stage to find more clues. Finally, I found another file file_3 hidden in Forensic_Encyption. We can modify the header back to PK and extract file_3.
file_3 is a pcap which recorded the traffic contained ESP protocol. We can decrypt the traffic with key.txt then get a html file.

Reflector:C Thin, beta, I, IV, II (T M J F), Plugboard: L-X/A-C/B-Y



The cipher is encrypted by enigma, but the website contained the encrypted key. Thus, we can decrypt the cipher easily.
I use this to encrypt and get the real flag.

Analysis-Offensive 200

cracktheflag.exe is a simple passcode validator which received a number and judge if the number is a valid passcode.
Surprisingly, this challege can be solved without any guessing.

The condition of the valid passcode is as below:

x1 = passcode / 10000 % 100
x2 = passcode / 100 % 100
x3 = passcode % 100

1. len(passcode) == 6 
2. `passcode` is primes
3. x1 is primes
4. x2 is primes
5. (x3 * x3 ^ x1) >> 8 == 0
6 sum(ascii(d) for d in passcode) - 288 is primes

At first, I tried to solve it with z3. However, it will spend a lot of time when checking prime. I decided to write a script to filter all possible solutions.
We can list all of the primes which satisified condition 1 and 2, then filter them with condition 3 to 6.
I found 7 solutions to satisify all conditions, and the biggest one is 236749:

  • 20509
  • 24109
  • 24709
  • 25309
  • 234149
  • 234749
  • 236749

The program description said there are 8 possible solutions. I have no idea where is wrong.
Anyway, the biggest passcode is the same.

flag: TMCTF{236749}

Forensic 100

The pcap is a DNS traffic. According the description, there are some messages hidden in the traffic. The hostnames are very suspicious because the last one is shorter than others. I concated them and decode it with base64, but getting nothing. I stuck in this stage until organizers posted a hint which said the cipher is base but not base64.
I tried to decode with familiar base familiy blindly, such as base128, base32. Obviously, it's wrong. Our teammate jeffxx found only 58 charcaters appeared in the cipher, then I tried base58 and success to decode the cipher. The plaintext is an article and the flag is at the end.

flag: TMCTF{DNSTunnelExfil}

MISC 100

I could not analyze the pcap with wireshark at first because the header was corruption. However, I saw there are some strings begin with CLIENT_RANDOM in the pcap. After googled, I known CLIENT_RANDOM is encrypted keys used in HTTP2 traffic. Thus, I tried to repair the pcap. file command said the pcap file is big-endien, but I compared with the other pacp file and found only the order of first 4 bytes is wrong. After fixed it, wireshark could open the pcap normally and I could dump the object in HTTP2 traffic manually. I'm not sure if the latest wireshark support to dump HTTP2 object.

The traffic is someone access a website about visual cryptgraphy. There are some pictures hidden in traffic and css. I stack at here then my teammate atdog found a methond to overlap the iamges and get the flag. <(_ _)>.

flag: TMCTF{CanYouSeeThis?}

comments powered by Disqus