Topic: In this hands-on workshop, you will be facing a variety of encryption binary. Your mission, if you accept it, will be to code in Python a corresponding decryptor.

We are going to discover some cryptographic algorithm implementation (including RC4, AES, Salsa20 and Chacha20), some hashing function (including MD5, SHA-1 and SHA256).
If you don’t know these algorithms, then this workshop is for you

Community Lab Session: This is not a training, it's a lab where hackers together explore how to break AI tool guard rails to produce a working exploit tool. It also includes a competition to see who can come up with the best working exploit for a minimally described advisory. Needs power, network, tables, chairs, radio mic.

Weds community training lab. David Jacoby + 1 Helper

Blockchain systems have redefined what it means for an attack to be permanent. Smart contracts are immutable by design; once deployed, their logic cannot be patched. Every vulnerability is forever, every stolen asset is gone, and the code is public for anyone to study and exploit.

Despite governing hundreds of billions of dollars, these systems keep failing in ways that any security practitioner will recognize: logic bugs that trust the wrong sequence of operations, access control that collapses under social engineering, and infrastructure implants so well-engineered they remain invisible for weeks.

In this talk, we examine three landmark exploits through the lens of the Die Hard films. This is not a gimmick; each film captures something structurally true about the corresponding attack. Gruber exploiting Nakatomi's own vault logic. Terrorists who seized the control tower, not the planes. Simon Gruber keeping everyone chasing riddles while the real crew drills into the Federal Reserve. The analogy makes unfamiliar attack surfaces immediately intuitive for any security practitioner, regardless of blockchain background.

We cover the 2016 DAO hack ($60M), the Parity and Ronin multisig failures ($905M combined), and the 2025 CPIMP attack, a clandestine supply-chain implant across seven blockchains, invisible to every monitoring tool, defused in a 36-hour coordinated war room. Three failure classes: trust in your code, trust in your keyholders, trust in your infrastructure.

Escape rooms are built to be escaped. One team, sixty minutes, a sequence of
locked doors where the key to each is hidden in the room before it.

Your delivery pipeline was not built to be escaped. It just is.

This talk walks a single unbroken chain from "ordinary employee with a source
control account" to "organization administrator of the entire cloud estate",
without a single CVE, exploit, or memory corruption bug. Every door is a
configuration someone chose deliberately, for a defensible reason, in isolation.
Eleven of them in a row is a catastrophe.

We pick the locks in order. Default write access nobody audits. A Terraform plan
that runs attacker code before a human reviews anything. An environment gate that
turns out to be a kitchen timer. A federation trust that checks the wrong claim.
An IAM edge that lets a service account promote itself. Then we run the whole
thing again as a speedrun, with the clock on screen.

Bring your own pipeline. You will want to go and check something afterwards.

Turning off TVs and monitors in public places is fun!
Learn to solder by making a kit that turns off these annoyances from 50 meters away.
For total beginners.
Enjoy life, and turn off a TV or two today!

TV-B-Gone is an open hardware remote control that can turn off any remotely controllable monitor in public places (or anywhere)! From across the street, through windows, in restaurants, bars, schools, airports... The new V2 of the kit makes it super easy to learn to solder, as well as easy to hack on.

Community Training Itinerary:

• Intro to remote controls   
• Learn to solder by making your own TV-B-Gone, step by step   
• Target practice is available all over the world after the workshop   

Taught by Mitch Altman.

Materials cost:
The workshop is free, but if you would like to partake in the hands-on aspects of the workshop, Mitch will have materials for €20.

Duration:
This community training takes 2 hours.

DICE-like measured boot guarantees software integrity bound to a
specific hardware even on low-powered devices with no Trusted
Execution Environment (TEE). An inherent problem in DICE-like measured
boot is that the first mutable code cannot be updated without losing
generated cryptographic keys.

In this talk I both introduce how we implemented DICE-like measured
boot in an open source project and how we solved updating device
applications without losing keys.

Agent2Shell is a family of zero-day RCE chains against the three flagship AI coding agents (Anthropic Claude Code, Cursor, and Google Gemini), all reproducible on the latest stable builds with default settings on Windows. The chains have been responsibly disclosed to the affected vendors and are under coordinated embargo. HITCON 2026 will be the first comprehensive technical deep-dive on the full research set.

What separates this work from the LLM-security research dominating the field today: the LLM is never invoked. Every chain triggers before any prompt reaches the model. Prompt Injection defenses, Guardrails, Output Filtering, and Agent Alignment all sit on the wrong side of the boundary and never observe the payload. We name this attack class Pre-Prompt RCE.

Several chains are 0-click: open a folder, double-click a small file, or run a routine command, and arbitrary code executes on a fully patched Windows machine. The remainder require at most one user action. Several are bypasses of recently patched CVEs in the same products, evidence the patches addressed instances rather than the underlying class.

The talk reconstructs the trust-boundary, auto-approval, and config-pollution failure classes shared across all three vendors, and shows why the bug class will keep producing instances until vendors redesign rather than patch. Live demos reproduce the most striking 0-click and 1-click chains end-to-end on stage.

Come and play some fun Hacker Jeopardy!

 _  _   __    ___  __ _  ____  ____      __  ____  __  ____   __   ____  ____  _  _  _   
/ )( \ / _\  / __)(  / )(  __)(  _ \   _(  )(  __)/  \(  _ \ / _\ (  _ \(    \( \/ )/ \  
) __ (/    \( (__  )  (  ) _)  )   /  / \) \ ) _)(  O )) __//    \ )   / ) D ( )  / \_/  
\_)(_/\_/\_/ \___)(__\_)(____)(__\_)  \____/(____)\__/(__)  \_/\_/(__\_)(____/(__/  (_)  

3rd party attacks happen all the time now, its mainstream. It has also hit Sweden with catastrofic outcome, and there is more to come. Today we will dig deep into the Public sector of Sweden, and their vendors. We have mapped every invoice to every municipality and found those that pose a high to critical risk towards Sweden and the municipalities. We will cover the data, the vulnerabilities and they dialogs between us, vendors, municipalities and agencies; With one goal in mind: What can we do to help out?

Learn to solder by making a way-cool, powerful music synthesizer, and
learn how to make cool music, sound, (and noise!) with computer chips
(The fancy word for making sound with a computer chip is Digital Signal Processing or DSP).

The SEC-T Music Synthesizer Badge kit (based on Mitch’s ArduTouch is an open hardware Arduino-compatible music synthesizer kit with a built-in Touch Keyboard, and with built-in speaker/amplifier. It is a really nice performing musical instrument.

This workshop is for total newbies to learn to solder.
This workshop is for total newbies to make their own SEC-T Music Synthesizer Badge and learn to make music, sound (and noise!) with computer chips.
Attendees take their completed synthesizer home at the end of the workshop.

The SEC-T Music Synthesizer comes pre-programmed with a way cool synthesizer. And Mitch will show you how to re-program it with other way cool (and totally different) synthesizers, and how to make your own synthesizers.

For ages 10 - 100.

Modern Wi-Fi chipsets are complex systems, but researching them often requires dealing with fragile vendor drivers, outdated kernel APIs and difficult debugging workflows. This talk presents a different approach: a complete implementation of the driver stack for the low-cost AicSemi aic8800 Wi-Fi 6 USB chipsets, first as a portable libusb userspace driver for both Linux and Windows, and then as a self-contained WebUSB application running directly inside the browser.

The project reconstructs the hardware initialization sequence, firmware loading process, USB transfer logic, receive and transmit paths and host-to-firmware command interface.
By moving the driver outside the kernel, the device becomes easier to instrument, fuzz and repurpose for offensive and defensive research, turning a cheap Wi-Fi dongle into a flexible platform for wireless reverse engineering and security research.

The second phase pushes the same model into the browser, through a retro terminal-style WebUSB interface, the user can interact directly with the hardware: send data to other clients, connect to networks, switch receive modes, enable raw frame injection, fuzz other network devices over the air and experiment with the runtime firmware rewriting feature for updating the firmware code during the driver execution.

We all know AI can find bugs. After a decade of research, I asked a harder question: can an autonomous system invent new attack techniques, and use them to hack live websites at scale? Building this sounded like a bad idea, so I did it.

It worked - I'll share an arsenal of new HTTP desync triggers, gadgets, and exploits that compromised banks, security solutions, and government infrastructure. Then I'll trace each discovery chain back through the HTTP Terminator, showing how to turn your personal expertise into an autonomous weapon - and the dark arts required to make it lethal.

I'll also share discoveries from beyond the autonomy horizon - some only reachable with a tight human/AI research loop, and others beyond AI's reach entirely. These include a powerful undisclosed recon technique, and anomalies that hint at new attack classes offering alternative paths to critical impact. I'll analyse the discovery process, sharing detailed experiments that probe the boundaries of what AI can and can't discover.

You'll leave with new exploits from desync triggers to undisclosed attack classes, and a blueprint for turning your instincts into an autonomous research cascade. And yes, I'll open-source the HTTP Terminator.

Have you ever imagined compromising almost every user of a major Linux distribution?

apt upgrade. apk add. dnf update. Every day, millions of servers execute these commands with absolute trust, relying on signatures to guarantee that an update is safe and official. But this only proves that a file successfully passed through a signing service at the very end of its journey. It does not validate the integrity of the massive, automated factory that built, processed, and advertised it. What happens when an attacker targets the distribution control plane itself?

Over several months, we audited the backend infrastructure powering major Linux distributions, including Ubuntu, Alpine, and Fedora. Moving beyond common supply chain threats, we focused on the internal systems that decide what becomes official. We discovered multiple independent compromise chains that allowed us to subvert the software supply chain entirely from the inside. By exploiting flawed trust boundaries in build artifact processing, over-privileged automation workflows, and logical gaps in metadata generation, we demonstrated how attackers can trick official infrastructure into blindly building, stamping, and delivering malicious updates to users.

This talk pulls back the curtain on the hidden complexity of modern software distribution. Join us to explore how backend pipelines can be compromised and weaponized.

The package was signed. The pipeline wasn't.

Care for a gaming break at the conference? Karategamers have brought their finest retro gaming consoles and equipment for you and your friends to relive or discover retro games from the the late 1900s and onwards. Still got the timing of the jumps or the order of the buttons drilled into your muscle memory? Find out at the Retro Arcade all afternoon, accompanied by Syntax Error DJ:s.
[The Arcade Is Open During Club SEC-T!!]

Learn everything you need to start making electronic projects!
You've probably heard lots about Arduino. But if you don't know what it is, or how you can use it to do all sorts of cool things, then this fun and easy workshop is for you.
As an example project, we'll be creating a TV-B-Gone remote control out of an Arduino you can take home with you.

Authentication reflection was thought to be solved after MS08-068, but recent research shows that Windows still exposes multiple ways to coerce systems into authenticating to themselves. This talk explores modern reflection techniques involving Kerberos, NTLM, and SPN manipulation, and how they can still lead to privilege escalation or even domain compromise. I will also present a critical reflection vulnerability in Windows Admin Center (WAC) discovered in July 2025 and fixed by Microsoft in early 2026.

Before the web. Before TCP/IP. Before “cloud.”
Some of the most powerful computers in the world were already online.
IBM mainframes didn’t grow up in the browser era. System/360 (1964), System/390 (introduced in 1990 and dominant throughout the 1990s), and today’s z/OS (released in 2000) were designed for batch jobs, green-screen terminals, and a world where the internet simply didn’t exist—yet these systems still quietly run banks, airlines, governments, and payment rails.

This talk is a guided tour of what happens when modern hackers bring cloud-era assumptions into a system that predates the web. We’ll break down how mainframes actually organize work (JES, JCL, RACF, CICS, PR/SM), why “root,” shells, and ports don’t mean what you think, and where attackers really interact with Big Iron today: transactions, protocols, management boundaries, and boring-looking misconfigurations that turn out to matter a lot.

No mainframe background required. No nostalgia cosplay needed. Just simple metaphors, diagrams, and stories from real red-team work—ending with a short checklist you can use the next time you find a system in scope that’s older than your favorite exploit class.

Since early 2025, software package maintainers have faced a sustained wave of attacks. Adversaries have used credential theft, exploitation of vulnerable CI/CD workflows, targeted social engineering, and adversary-in-the-middle (AiTM) phishing to compromise upstream packages. The damage is real: the Shai-Hulud worms, TeamPCP's compromise of Trivy and the LiteLLM PyPI package, and North Korea's takeover of the widely used Axios package. These attacks infected thousands of developers and pipelines.

This talk covers how we hunt for malicious and compromised packages at scale, give the community the right tools to do the same, and secure an internal engineering organization of over 3,000 engineers who install third-party packages every day.

HarmonyOS (HMOS) is a mobile platform developed by Huawei. It is rapidly gaining popularity in China, and Huawei are open about their ambitions for becoming a global contender to iOS and Android. While the Chinese hacking community has made good progress on finding ways to play with HMOS, it remains fairly unknown (and unhacked) in the West. This is not helped by the fact that the system's development tools are somewhat restricted outside of mainland China.

In this session, I will talk about how HMOS works, and why I find it interesting. If you bring a Windows or macOS machine, I will also help you set up a decent playground environment to start with. Our hope is to build a small community of geeks poking the system with a stick before it goes global (if it ever does)

In this keynote, we take a step back from the hype cycle and look at security through a longer, vintage, grumpy old hackers lens. From the early days of classic computer hacks to the evolution of modern vulnerability research, we revisit the techniques, mindsets, the breakthroughs and the people responsible for those, with a focus especially on the gaping holes that shaped the industry. Along the way, we’ll examine how each “revolution” -from automated scanners to current day AI- was (or is) supposed to change… everything.
Grumpy old hacker's voice: “And yet, the core of security has remained stubbornly human.”
Through real-world facepalming examples (here come the tears), historical parallels, and a candid look at today’s AI-driven slop, this talk challenges the narrative that AI is replacing hackers. And kittens? They’re probably fine. But just to be sure, bring napkins to this talk. Please.

I'm going to be taking a baseball bat to frontier LLMs, using math and statistics. I'll be showing logs of how they literallly talk themselves into lying to you. It's bad. Bring a helmet.

Prompt engineering has become harness engineering. Openclaw and now codex are storing local files and 'memories' to try and handle the 'context window problem'. Moltbook has 3 million 'agents'. Openclaw is being used as a c2 now.

The attack surface is growing so rapidly we can barely keep track of it. This talk will explore all this new attack surface, and cover some of the things you can do about it.

Welcome to Club SEC-T!

This is the official SEC-T party on Thursday night. To make this as much fun as possible, it's free to attend even if you don't have a conference ticket, and no registration is required either - simply drop in if you're 18 years or older 🙂

19:00 Doors open to the public
19:15 Wiklund live-on-stage
20:00 Cyber Security Quiz #20
21:30 MBR live-on-stage
22:45 Syntax Error DJs
01:00 Doors close

Club SEC-T is -free- to attend for anyone 18 years or older, whether you're a conference guest or just want a ridiculously geeky Thursday night and don't even know what SEC-T is!

This also means, as a conference guest, you're free to invite your non-ticket-holding friends.
No ticket nor registration needed. The schedule, once announced, is preliminary and subject to change.

Club SEC-T is also open on Thursday from lunch and on Friday until lunch for conference ticket holders with a more relaxed vibe, background music, Karategamers Retro Arcade and conference feed on the big screen..

Everyone's talking about AI defending the perimeter. This one's about the other side of the table. I'll show you how public AI models quietly became the most versatile thing in an attacker's kit, a real force multiplier for building exploits and offensive tooling at a speed that just wasn't possible two years ago.

We'll dig up old, half forgotten techniques and watch AI breathe new life into them, then go fully modern with a stack of live hacking demos. But this isn't just for show: you'll walk out with practical, take it home ways to start folding AI into your own workflow, including a run from a raw bug all the way to a working exploit. Minimal slides, maximum terminal. You'll leave knowing how cheap and fast offensive capability has gotten, and with a real starting point for doing it yourself.

Demos include SSH session hijacking, TTY/PTY injection, password stealing, command injection and old hacker tools on steroids.

Care for a gaming break at the conference? Karategamers have brought their finest retro gaming consoles and equipment for you and your friends to relive or discover retro games from the the late 1900s and onwards. Still got the timing of the jumps or the order of the buttons drilled into your muscle memory? Find out at the Retro Arcade all afternoon, accompanied by Syntax Error DJ:s.
[The Arcade Is Open During Club SEC-T!!]

Turning off TVs and monitors in public places is fun!
Learn to solder by making a kit that turns off these annoyances from 50 meters away.
For total beginners.
Enjoy life, and turn off a TV or two today!

TV-B-Gone is an open hardware remote control that can turn off any remotely controllable monitor in public places (or anywhere)! From across the street, through windows, in restaurants, bars, schools, airports... The new V2 of the kit makes it super easy to learn to solder, as well as easy to hack on.

Community Training Itinerary:

• Intro to remote controls   
• Learn to solder by making your own TV-B-Gone, step by step   
• Target practice is available all over the world after the workshop   

Taught by Mitch Altman.

Materials cost:
The workshop is free, but if you would like to partake in the hands-on aspects of the workshop, Mitch will have materials for €20.

Duration:
This community training takes 2 hours.

In our 2023 BSides London talk, we revealed a number of vulnerabilities in a popular line of smart padlocks which allowed unauthorised users to obtain unlock credentials for any affected lock and open it without the owner’s knowledge.

Building on this research, we've developed a training lab to help introduce newcomers to IoT and mobile hacking. Our lab setup replicates the app responsible for controlling the smart locks, and serves as a practical introduction to mobile application reverse-engineering, Bluetooth Low Energy communications, with a garnish of simple API testing. We'll bring the locks and mobile phones to control them.

Bring your own laptop with a Linux VM (Kali works). We'll bring the phones, the locks, and the app. :)

This talk covers how the Lovable security team identified and tracked a global attack attempted to compromise our users' Github projects with malicious code. We analysed millions of Lovable projects across hundreds of thousands of customers and tracked the campaign through its different phases meanwhile we built tools to intercept and prevent the attack.

The attacker used a mix of human and non-human identities, introduced malicious code through direct commits and merge paths, and bypassed expected platform provenance. We observed payloads hidden in frontend configuration files and used obfuscated staged JavaScript to fetch and execute encrypted second-stage code. The attack affected anyone who worked with or built the code.

This talk shows how to detect and contain this pattern by correlating Git events, CI/CD metadata, and platform edit telemetry. Attendees leave with knowledge of how to prevent similar attacks from happening, deploy detections, triage logic, and conduct incident response.

The Cult Of The Dead Cow (cDc) is breaking the internet with Veilid, an open-source, peer to peer, mobile-first, network application framework. Veilid goes above and beyond existing privacy technologies and has the potential to completely change the way people use the Internet. Veilid has no profit motive, which puts it in a unique position to promote ideals without the compromise of capitalism. With Veilid, the user is in control, in a way that is approachable and friendly, regardless of technical ability.

This framework shares some similarities with IPFS and Tor in its overall design, but it is built to deliver better performance while natively supporting all services through a privately routed network. It allows developers to create fully decentralized applications without depending on a blockchain or transaction-processing layer as the foundation. The framework can either be integrated directly into end-user applications or operated as a standalone headless node by advanced users who want to contribute resources and help strengthen the network. Thousands have already joined the Veilid network, integrating it into their daily lives and building privacy-focused projects on top of it.

Together, we can give the world the private Internet we should have had all along.

I will talk about how you yourself can build a drone, what a drone even is, how it's used by both civilian and military. Stockholm Tactical Fundraising Ukraine (STFU), also known as Tacticats, is a volunteer based organisation (idéell förening), our purpose is to fundraise, buy equipment and deliver it to Ukrainian defenders. We work with targeted aid, responding to specific requests from people at or near the frontline.
We also organize workshops where participants can learn how to build, or fly fpv drones. The built drones are then sent to Ukraine. So far, we've built around 60 drones since we started hosting these workshops back in November 2025.

When you think of hacking browsers, you perhaps think of V8 heap exploitation, deep-dive fuzzing, crazy sandbox escapes, and so on. But what if I told you that you can still find vulnerabilities in major browsers that don’t require any technical knowledge? Bugs you can even run into by accident!
In this talk, I’ll take you through my journey of how I “accidentally“ found a vulnerability in Google Chrome. And how that led me to find more vulnerabilities in Chrome as well as some vulnerabilities in Mozilla Firefox and many more bugs in other products.
So if you’re keen to find out how I could, with minimal user-interaction, steal your private GitHub repositories, then this talk is for you!

Learn to solder by making a way-cool, powerful music synthesizer, and
learn how to make cool music, sound, (and noise!) with computer chips
(The fancy word for making sound with a computer chip is Digital Signal Processing or DSP).

The SEC-T Music Synthesizer Badge kit (based on Mitch’s ArduTouch is an open hardware Arduino-compatible music synthesizer kit with a built-in Touch Keyboard, and with built-in speaker/amplifier. It is a really nice performing musical instrument.

This workshop is for total newbies to learn to solder.
This workshop is for total newbies to make their own SEC-T Music Synthesizer Badge and learn to make music, sound (and noise!) with computer chips.
Attendees take their completed synthesizer home at the end of the workshop.

The SEC-T Music Synthesizer comes pre-programmed with a way cool synthesizer. And Mitch will show you how to re-program it with other way cool (and totally different) synthesizers, and how to make your own synthesizers.

For ages 10 - 100.

You don't realize it until you see them; they're everywhere. From Wireless ISP links to the frontline of modern warfare. But nobody found anything?
The devices behind those links are Ubiquiti AirMAX: critical infrastructure on a 17-year-old Linux kernel and a custom 802.11 extension built on "security by obscurity."
So we took it apart. This talk covers our reverse engineering of the AirMAX protocol, AirOS, and the kernel modules behind this proprietary mode. It rides on 802.11 Information Elements that look encrypted, but we'll show why they aren't.
What we found: two critical vulnerabilities (CVE-2026-21639, CVE-2026-21638) across airMAX AC, airMAX M, airFiber, and GigaBeam, over 50 devices. These are the bugs from the movies: Over-The-Air, unauthenticated, kernel-privilege RCE. No network access, just line of sight. They affect every AirMAX device ever shipped.
We disclosed them through Ubiquiti's bug bounty program. The bugs were rated "Adjacent", except adjacent here means kilometers away.
The same hardware can be turned around and pointed at the problem: we'll repurpose these devices as recon tools and release open-source software to locate AirMAX networks in the wild.
This talk is about our journey, our tooling, and the state of security.

Ever wondered what's going on inside your iPhone?
It's an incredible can of worms, that Etum/Agnes would love to explore with you!

We'll crack open how iOS and its apps actually work under the hood, then get you hands-on with Arm64 and LLDB ( the real tools that jailbreakers and security researchers use) before setting you loose on live disassembly and debugging. Bring questions, a linux laptop and a can do attitude!

This workshop is beginner-friendly, no experience required (but if you've touched a terminal before, you'll feel right at home)

Workshop outline:

• Introductory talk about how iOS/iOS apps work "under the hood".
• Arm64 and LLDB basics, follow-along.
• Hands-on simple disassembly and LLDB debugging.
• Post-workshop reversing material
• ... Probably low-level iOS tangents if you poke Etum enough about it

Materials:
Bring your own laptop. (Mac or Linux)
For the hands-on section, we will supply as many Jailbroken iOS devices as we have available; otherwise, go ahead and bring a macbook with vphone-CLI [https://github.com/Lakr233/vphone-cli] installed (Requires AMFI turned off over csrutil)

Modern Linux systems run a long chain of code before the kernel takes over - firmware, bootloaders like GRUB, and early runtime environments. Although each component is well understood, the transitions between them are often treated as simple handoffs, obscuring key assumptions about what state persists, what gets reinitialized, and where execution can truly be intercepted.

This talk offers a practical examination of one of the most important transitions: the GRUB-to-kernel handoff. Through reverse engineering GRUB and analyzing how the kernel image is loaded, decompressed, and executed, it shows how early-boot control flow is assembled - and where it can be influenced reliably.

Instead of staying theoretical, the presentation focuses on real constraints that shape boot-chain abuse. It explores which execution contexts survive across stages, which don’t, and how that affects the feasibility of establishing persistent control before OS security mechanisms activate.

To ground the discussion, a UEFI bootkit is built within the reconstructed execution flow. The implementation demonstrates concrete ways to subvert the Linux boot process, along with limitations and failure modes that appear when early-execution assumptions break.

Overall, the talk clarifies how early boot works in practice: what can be controlled, what cannot, and what that implies for both offensive techniques and defensive visibility during one of the least observed phases of system execution.

Although more than eight years have passed since transient execution attacks such as Spectre and Meltdown shocked the computer industry, this class of vulnerabilities remains largely unexplored beyond well-studied targets such as the KVM hypervisor. Less mainstream, yet widely used, hypervisors have received comparatively little scrutiny from security researchers investigating transient execution attacks. While we often assume that our virtualization platform provides a secure and isolated environment for cloud workloads, malware analysis, and other untrusted software, that assumption does not always hold.

In this talk, we take a deep dive into transient execution attacks, with a particular focus on Branch Target Injection (BTI)-based Spectre attacks. After explaining the underlying exploitation techniques, we examine the mitigations implemented by modern hypervisors, using KVM as a reference point. We then turn our attention to FreeBSD's bhyve and Oracle VirtualBox and show that neither provides adequate protection against BTI attacks. To demonstrate the practical impact, we present end-to-end proof-of-concept exploits against both platforms that leak arbitrary memory from the host.