←
FuzzingLabs830 posts
FuzzingLabs avatar

FuzzingLabs

@FuzzingLabs

Research-oriented Cybersecurity startup specializing in #fuzzing, Vulnerability Research & Offensive security on Mobile, Browser, AI/LLM, Network & Blockchain.

◷ Joined August 2020▧ 305 Photos and videos
FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·30 Oct 2025

💥 We’ve just raised €1M in pre-seed funding to accelerate the development of FuzzForge. When I started FuzzingLabs, everything was bootstrapped: our audits, our trainings, our R&D. No investors, no funding. Just a passionate team obsessed with offensive security and the belief that we could build something different. Three years later, we’re 30 and we are now entering a new chapter. This funding will allow us to: - accelerate the open-source development of FuzzForge, - build its marketplace of agents and workflows, - and expand the SaaS version to automate vulnerability research at scale. A huge thanks to @class_lambda and @ergodicgroup for their strategic support and trust in our vision: --> making offensive security more intelligent, collaborative, and automated. FuzzForge is already open source and under active development. You can check it out here: 🔗 github.com/FuzzingLabs/fuzzf…

Post attachment
Patrick Ventuzelo avatar
⟳ FuzzingLabs retweeted
Patrick Ventuzelo@Pat_Ventuzelo·Oct 28

Project Zero (@5aelo) pitted an LLM agent (BigSleep) against a fuzzer (Fuzzilli) on V8: 45 vs 37 vulns, ~25% overlap. -> Bugs hard for fuzzers are hard for agents too. -> AI won't kill fuzzing. My take long term: Agents that drive fuzzers will win. projectzero.google/2026/10/f…

Post attachmentPost attachment
FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·Oct 8

First attempt, first win. 🎯 Congrats to @LucasVanHaaren and @0xNaSu for popping the Brother MFC-L8970CDW with a single 0-day. Thanks @thezdi for a great event! #Pwn2Own

Brother what? Brother exploited! Lucas Van Haaren (vhash, @LucasVanHaaren) and Hugo Leclercq (0xnasu) of FuzzingLabs (@FuzzingLabs) exploited Brother MFC-L8970CDW with a single zero-day earning $20,000 and 2 Master of Pwn points #Pwn2Own

Post attachment
FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·Sep 8

With the harness up, the first crash came quickly. KASAN report in batadv_iv_ogm_send_to_if, the function that splits aggregated OGMs out of a forward packet.

Post attachment
FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·Sep 8

One blocker: no Syzlang grammar existed for batman-adv packets. So we wrote one. All 13 packet types (OGM, OGM2, ELP, ICMP, TVLV, fragmented unicast )derived from: include/uapi/linux/batadv_packet.h into sys/linux/vnet.txt.

Post attachment
FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·Sep 8

The hard part of kernel fuzzing is never the fuzzer. It's reaching the code. We used TUN/TAP to inject raw Ethernet frames directly into the VM's network stack, through Syzkaller's syz_emit_ethernet pseudo-syscall. Approach borrowed from @andreyknvl write-up.

Post attachment
FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·Sep 8

B.A.T.M.A.N. "Better Approach To Mobile Ad hoc Networking" is an L2 mesh routing protocol from the Freifunk community, shipped as a kernel module. Think of it as one big switch where every device is a peer. Originator Messages and ELP keep the routing state alive.

Post attachment
FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·Sep 8

A file in the Linux kernel with 0% Syzbot coverage is an invitation... That's how our team at @fuzzinglabs ended up looking in batman-adv, the mesh networking subsystem (B.A.T.M.A.N. = Better Approach To Mobile Ad hoc Networking). Nobody had ever fuzzed the Linux kernel's mesh networking subsystem from the frame reception side. So we did. Three bugs, one patch now merged in mainline. 🧵

Post attachmentPost attachment
FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·Jul 2

Cryptography on embedded devices: picking AES vs Ascon is the easy part. The hard part? Where keys live, whether the chip has any trustworthy RNG, if it can persist a nonce counter, and which accelerator you're forced to use. Our 101 guide 👇 fuzzinglabs.com/crypto-embed…

FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·Jun 30

Our CEO @Pat_Ventuzelo opened @_leHACK_ 2026, France's biggest hacking con, with his keynote: "No need to be a Mythos to do offensive security." You don't need a classified frontier model. The moat moved from model access to the harness and the building blocks are already in your hands. Slides 👇 You don't have to be a Mythos. You just have to keep building. fuzzinglabs.com/wp-content/u…

Post attachment
FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·Jun 16

🛰️ We broke RTKLIB, the open-source engine behind centimeter-accurate GPS in drones, boats, autonomous vehicles & survey networks. 4 memory-corruption bugs in its RTCM3 / RINEX decoders. No auth. Reachable from a single crafted correction stream or one booby-trapped file. 1 OOB write + 3 OOB reads. A rogue NTRIP caster or a MITM'd stream is enough to crash or corrupt the positioning stack of a machine that moves itself. Reported upstream, coordinated disclosure (RTKLIB #796–799). Full write-up 👇 fuzzinglabs.com/breaking-rtk…

Post attachment
FuzzingLabs avatar
FuzzingLabs@FuzzingLabs·Jun 2

Excited to be there end of the year !! Looking forward to show to the community what we have build with fuzzforge, our ai agents orchestration platform for embedded security !

Excited to welcome @FuzzingLabs as an Exhibit Sponsor for Hardwear.io Netherlands 2026. 🚀 Looking forward to having their team and research expertise. Registrations & CFPs now live: hardwear.io/nl-2026/cfp/?utm… #HardwearNL2026 #ExhibitSponsor

Post attachment