RewriteLab (@RewriteLab) profile picture

A specialized research team focused on web security vulnerabilities and exploitation techniques

Joined July 2025
Photo by RewriteLab (@RewriteLab)Photo by RewriteLab (@RewriteLab)Photo by RewriteLab (@RewriteLab)Photo by RewriteLab (@RewriteLab)Photo by RewriteLab (@RewriteLab)Photo by RewriteLab (@RewriteLab)Photo by RewriteLab (@RewriteLab)
We published a new research article: The Frenemies, GC and JIT! In this article, we dive into CVE-2026-112653, a JavaScriptCore DFG bug caused by an unexpected GC safepoint in the NewArrayBuffer allocation path. This breaks the epoch invariant used by StoreBarrierInsertionPhase, causing a required write barrier to be incorrectly elided. We walk through the bug from JIT IR and write-barrier analysis to reliable exploitation: racing JavaScriptCore’s concurrent marker, creating an untracked black-to-white edge, turning it into a GC-induced use-after-free, and ultimately building heap read/write primitives in the WebContent process. We also cover the surprisingly small one-line patch that fixes it. Huge thanks to our team member @m411k_ for conducting this research! Check out the PoC! Full article: rewritelab.org/researchs/12/
2
22
3
116
8,879
RewriteLab retweeted
1/16 I believe there are people facing the same concerns as our team, so I want to share this publicly. Our team recently went through a major period of stagnation. The core problem was that fewer people wanted to participate in technical research.
3
5
21
2,179
Our long-standing RewriteLab research pages have been completely redesigned! A huge thanks to @OpenAI for helping with the design! ❤️ You can now check out the new look at: rewritelab.org/research Stay tuned for the new research we'll be publishing in the future as well : )
8
1
30
1,823
🚨 Bounty Update Someone solved the challenge in just 58 minutes using an unintended solution. (It was a completely unexpected approach—seriously impressive!) However, they kindly said the bounty should go to the first person who solves it via the intended path, and voluntarily declined the reward. (Huge thanks for the incredible sportsmanship!) The bounty is still open. 🎉 That said, I'm adding one new rule: ✅ The challenge must be solved via the intended solution path. (The current unintended path will be patched soon.) If you solve the challenge, please open a ticket and explain your solution. If it matches the intended path, you'll receive the bounty.
🚨 OPEN BOUNTY — $300 Can your AI agent solve an unsolved CTF challenge? We launched the damn-vulnerable-web on February 1st, and one challenge has remained at 0 solves for over 5 months despite the rise of LLM slopping. This bounty is only available until the end of July. Challenge: chronostasis First Blood: 200 USDT 300 USDT if you solve it using a fully autonomous LLM agent ("pure slopping") and can demonstrate the complete solving process. Still sitting at 0 solves. Huge respect to @icesfont2 for creating such an insane challenge 👏 Bring your best agent. Let it cook!
4
2
51
8,077
Replying to @morbintimenfts
unfortunately, a Discord account is required
1
1,850
You can try the challenge here : wargame.rewritelab.org
7
2,875
🚨 OPEN BOUNTY — $300 Can your AI agent solve an unsolved CTF challenge? We launched the damn-vulnerable-web on February 1st, and one challenge has remained at 0 solves for over 5 months despite the rise of LLM slopping. This bounty is only available until the end of July. Challenge: chronostasis First Blood: 200 USDT 300 USDT if you solve it using a fully autonomous LLM agent ("pure slopping") and can demonstrate the complete solving process. Still sitting at 0 solves. Huge respect to @icesfont2 for creating such an insane challenge 👏 Bring your best agent. Let it cook!
11
18
2
217
30,940
We published a new research article on the Chromium 146 Renderer Process! In this article, we start from the CVE-2026-3910 Maglev write barrier elision bug and walk through the full exploit chain: building a V8 heap R/W primitive via a GC-induced UAF, achieving an out-of-sandbox read using WebAssembly internals, abusing JSPI UAF and StackMemory / JumpBuffer, and ultimately reaching renderer process RCE. Our goal was to provide a structured explanation of how modern V8 exploitation works in practice, from compiler-level bug analysis to sandbox-boundary primitives and final code execution. Huge thanks to our team member @m411k_ for conducting this research! Check out the PoC! Full article: research.rewritelab.org/2026…
1
34
169
12,973
Replying to @PallavTrip9276
Thank you for your interest in our research I do not believe permission segmentation and HITL can prevent prompt injection attacks themselves. However, they are necessary risk management strategies for limiting the impact when a prompt injection attack occurs
15
We also discuss mitigation approaches including permission segmentation, sandboxing, validation layers, and human-in-the-loop design
1
1
294
This write-up also includes several case studies, such as WhatsApp MCP, GitHub MCP, and OpenClaw, to show how these risks can appear in real systems
1
1
370
One point we wanted to highlight is that prompt injection in agentic systems should be examined not only at the prompt level, but also at the levels of memory, tool usage, permissions, and system architecture
1
1
143
We published a new research article on prompt injection in modern agentic systems This write-up covers: - direct and indirect prompt injection - multi-turn attack methodology - tool-calling and MCP-related abuse cases - case studies including WhatsApp MCP, GitHub MCP, and OpenClaw mitigation strategies such as permission segmentation, - sandboxing, PTC, and HITL Our goal was to provide a structured overview of how these attacks work and how they can be addressed in practice Full write-up: research.rewritelab.org/2026…
2
9
26
2,048
D-10 !!!!!
[ RewriteLab Web Security Research Team, 2026 First Half Researcher Recruitment ] Rewrite is a specialized web security research team composed of web hackers from around the world. Researchers from various regions including Korea, Europe, Asia, and Africa collaborate to conduct in-depth research on the latest web exploitation techniques and technologies, while also working on a range of web security related projects We are now publicly recruiting new researchers who would like to join RewriteLab and conduct research together with us For detailed information about the recruitment requirements and the application process, please refer to the recruitment page below! recruit.rewritelab.org
1
10
2,266
We have successfully published a new research article! This research takes an in-depth look at several interesting security incidents that occurred in 2025 and analyzes them in detail While some of these incidents were already widely known, this research focuses more closely on cases that people may have only glanced over without examining thoroughly Special thanks to One, TCP/IP, and @filime_sec for conducting this research! We hope it receives a lot of interest! : ) research.rewritelab.org/2026…
9
20
1,851
[ RewriteLab Web Security Research Team, 2026 First Half Researcher Recruitment ] Rewrite is a specialized web security research team composed of web hackers from around the world. Researchers from various regions including Korea, Europe, Asia, and Africa collaborate to conduct in-depth research on the latest web exploitation techniques and technologies, while also working on a range of web security related projects We are now publicly recruiting new researchers who would like to join RewriteLab and conduct research together with us For detailed information about the recruitment requirements and the application process, please refer to the recruitment page below! recruit.rewritelab.org
13
2
70
11,761
Hello! We’ve just launched a new wargame site called damn vulnerable web! It consists only of web challenges, primarily designed for intermediate to advanced players rather than beginners. We hope this wargame helps more people gain deeper and broader knowledge in web hacking :) For now, we’re planning to accept only 300 users initially for open beta testing and capacity checks. Starting from this tweet, we’ll gradually increase the number of allowed sign-ups each week. Your interest and support will be a huge help to our future activities We’ll do our best to deliver even better work going forward. Thank you! Wargame site: wargame.rewritelab.org Join our Discord: discord.gg/wYAm2n4M4J
6
90
5
511
28,888