Sunshine CTF 2025 - AstroJIT AI (pwn)
This pwn challenge requires playing with the input of weights and biases of an AI chatbot to leak the flag.
This pwn challenge requires playing with the input of weights and biases of an AI chatbot to leak the flag.
The challenge required us to find a 10-float "genetic code" that maximizes a hidden "survival rating" function to an average of 95.0% within 100 generations. This is solved using a Genetic Algorithm, an optimization technique that mimics natural selection.
This attack is a classic cryptographic vulnerability called CBC Bit-Flipping Attack, which targets data encrypted using the CBC mode. The goal of the challenge is to get the server to decrypt a malicious packet that results in a specific, secret value for the `device_id` (0xdeadbabe).
The core problem is that the same secret message (P) is being encrypted repeatedly with a predictable, but incrementing, counter. Because the message is longer than the known plaintext, we must retrieve two consecutive ciphertexts (C0 and C1) to piece together the full encryption keystream (K0).
This is a writeup for the phished challenge for K17 CTF. Challenge Details The challenge description reads: We fired Billy last week after he failed a phishing test for the 6th time. We wiped his machine, but now we really need one of the files that was on it. Maybe
Unleash the Kraken
This post is a write-up for the pwn.hateful challenge in Nullcon Goa HackIM 2025 CTF. root@72f9eb9e3ebc:/chal/NULLCON/hateful# ./ld-linux-x86-64.so.2 --library-path . ./hateful My Boss is EVIL!!! I hate my Boss!!! These are things you really want to say to your Boss don't you? well
Subscribe to stay up to date on OSIRIS events and secrets đź‘€
This post is a write-up for the pwn.hateful2 challenge in Nullcon Goa HackIM 2025 CTF. root@72f9eb9e3ebc:/chal/NULLCON/hateful2# ./ld-linux-x86-64.so.2 --library-path . ./hateful2 _______ _________ _______ _______ _ _______ |\ /|( ___ )\__ __/( ____ \( ____ \|\ /|( \ / ___ ) | ) ( || ( ) | ) ( | ( \/| ( \/| ) ( || ( \/ ) | | (___) || (___) | | | | (__ | (__ | | | || | / ) | ___ || ___ | | |
By Smallfoot Feb 1, 2025
A CSAW '24 Quals challenge
Maybe next time you should record your music in an acoustically treated space by slimecat Plug This challenge was written for CSAW 2024 qualifiers by a member of NYU's OSIRIS Lab. CSAW is the world's most comprehensive student-run cybersecurity event and it's hosted here
This was the easiest rev challenge in brics+ CTF, hosted by ITMO in Russia with the support of several (as of the time of writing) US sanctioned Russian companies. The challenge includes two files, exfilter.ko, a kernel module, and exfilter_traff.pcapng, a network capture. Opening exfilter_traff.pcapng
This is the first, and easiest challenge in the brics+ CTF, hosted by ITMO in Russia with the support of several (as of the time of writing) US sanctioned Russian companies. This challenge suggests that you have to send a tiny ELF x64 executable which calls /bin/sh or equivalent.
shellcoding using hash functions! hash it 0 The hash-it-0 challenge is from DEFCON’s 2022 qualification round. It is like a normal shellcoding challenge but with mild steroids… The challenge involves shellcoding and we need to encode the shellcode for the target program to process and execute. Analysis The challenge
By applying traditional fuzzing techniques, we achieved high throughput SMT constraint solving. We were able to achieve 23 billion execs/s using GPU acceleration. SMT solvers typically apply very complicated algorithms to determine if a set of constraints is satisfiable (and produce a solution). The approach that we have taken
This year’s CSAW quals was the first time I authored a challenge, the challenge is called krakme and it was a 200pt rev challenge. You can view the challenge files in the repo here. The idea for this challenge was formed after reading this paper The impact of GPU-assisted
CTF Writeup for DawgCTF 2020: Where we roppin boys? DawgCTF 2020: Where we roppin boys? (350 points) * Challenge author: trashcanna Challenge files can be found here. Overview The rop program we are presented with is quite simple. main runs the welcome function, then the tryme function. The welcome function mmaps
Vasilisk is a V8 JIT fuzzer that focuses on optimization passes. Background Fuzzing consists of randomized input generation to automate the process of finding crashes in programs. The probability of finding a crash increases as code coverage in the application increases. In a well formed fuzzing campaign, test cases are
This post is on how the Ghidra decompiler works, and how to make it work for Go. Part 1 At the time of writing, the most recent major version of Go is Go 1.13 golang.org The most recent major version of Ghidra is Ghidra 9.11 github.com/