**Digital technology has made our lives brighter, more convenient, and more efficient. However, the rise of sophisticated cyberattacks is undermining people's confidence in this brave new cyberworld. The Positive Hack Days 12 international festival on practical security, which will be held on May 19 and 20 in Moscow's Gorky Park, will reinstall this confidence by opening its doors to people from outside the cybersecurity community for the first time. The new cyberfestival format will be of interest not only to cybersecurity specialists, government, and businesses, but also to ordinary people who will get to see how the digital world truly works and take a step towards trusting new technology.**

This year [PHDays](https://phdays.com/en/) will no longer be a closed event, and the new venue was not chosen accidentally. Known for its festivals and events, Gorky Park is one of the most famous public spaces of the capital, visited by about 30 million people annually. For the two days in May, the park will be changed beyond recognition. The forum will be divided into two zones: the central part of the park will be turned into an open digital city where everyone will be able to learn the basics of cybersecurity, while on the Pushkin embankment, a cybervillage will be built where the usual forum program with reports, discussions, and competitions will take place.

The open area of the forum will represent a modern metropolis hit by an onslaught of cyberattacks. Guests will take on the role of security researchers and join in a quest in which they will try to find and fix vulnerabilities in various infrastructure components of the digital city. For this, they will need to use a special ability—to see the city through the eyes of hackers.

In the cybervillage, digital nomads will be spread out amongst the tents. The elite of the cybersecurity industry will also gather there, to discuss the most acute issues related to cybersecurity. This year, the forum will see more than 100 speakers, including developers, security experts, hackers, CIOs, and CISOs of large IT companies.

The PHDays 12 technical agenda covers such issues as:

- Countering hackers and responding to incidents
- New attack techniques and vectors, exploitation of vulnerabilities
- Application development and security
- ML technologies and cybersecurity

The first speakers are already confirmed. They include Wildberries Information Security Director Anton Zhabolenko and Head of Infrastructure Security Pavel Parkhomets, who will share their experience in designing a privileged access management (PAM) system for a distributed Linux infrastructure; Oleg Skulkin, Head of Cyberthreat Intelligence at BI.ZONE, who will talk about how, why and by whom attacks are made on Linux-based infrastructure amidst import substitution; Alexander Korotin, Senior Application Security Specialist at Kaspersky, who will discuss the vulnerability of backup tools which can become an entry point for attackers striving to compromise the network; and Sergey Prilutsky, Chief Research Officer at MixBytes, who will break down common vulnerabilities of Ethereum smart contracts and explain how to protect your protocols from new attack vectors.

Security researchers still have one last opportunity to apply for a speaker slot in the cybervillage: [Call For Papers has been extended through April 7.](https://cfp.phdays.com/login?utm_source=phd&utm_medium=cpm&utm_campaign=phd11&utm_content=main-banner)

The Standoff cyberrange is a model of a digital state with three dozen teams of attackers and defenders battling for control over its resources. This time, the cyberrange will be modernized and expanded. In addition to the energy, oil and steel sectors, the nuclear industry sector will be built. New facilities will appear in all sectors, such as a solar power plant and a dispatch control center in the power sector. The attackers will try to trigger more than 100 non-tolerable events.

For two days, the PHDays 12's live studio will cover the key events of the forum and the Standoff cyberbattle, and host discussions on the most burning cybersecurity issues with the most prominent speakers and guests.

On the forum's stage, information security specialists will explain the basics of cybersecurity to everyday people who don't otherwise think much about it.

Please note that those wishing to attend the closed part of the forum need to [purchase a ticket.](https://phdays23.ruwidgets.ru/en/) This can be done right now: up to midnight on April 7, the ticket to the cybervillage will cost only 14,700 rubles; from April 8, the price will be 21,300 rubles. The central part of the park with the digital city will be open and free to all visitors.

See you on May 19 and 20 at PHDays in Gorky Park!

PHDays 12: trusting the tech thanks to cybersecurity

## [Call for speakers is now open.](https://cfp.phdays.com/login)

Following last year's cyberattacks, even previously laid-back people are now concerned about cybersecurity. That's why we decided to make the twelfth Positive Hack Days information security forum even more open and large-scale: it will be held at Gorky Park in Moscow on May 19 and 20, 2023. We've already started preparing for the event and are accepting applications from speakers. You can submit a talk topic until March 25.

In 2022, PHDays set another attendance record: 10,000 visitors came in person and over 130,000 viewers watched online. In 2023, the forum's audience will expand even further thanks to the city festival format and the open-air venue in one of the capital’s most popular parks.

If you still think that cybercriminals can't be interested in you or your company, this is a delusion with potentially disastrous consequences. Information security concerns absolutely everyone and should be common public knowledge. That's why this year's forum will have a more open format, with a range of activities for everyone to get involved in. At the same time, all the traditional events and tracks, including cybersecurity competitions, will remain.

PHDays will once again host the Standoff cyberbattle in which attackers and defenders will test the infrastructural security of a digital state and uncover possible chains of non-tolerable events in various industries. If you want to share your own valuable discoveries and developments, talk about bug bounty, participating in Standoff, or what it's like to be a red or blue team expert, become a speaker!

The forum is open to all researchers: the program committee will consider applications from both established infosec experts and newcomers. More than anything, we'd love to hear original perspectives on current problems in information security.

At PHDays 12, we're looking forward to hearing expert talks on both defensive and offensive security. We are also looking for speakers on the topics of development, blockchain, and machine learning in information security. But if you have an interesting cybersecurity talk or presentation on a different theme, we'll be happy to consider it.

If you have something to say about data leaks, turnover penalties, mergers and acquisitions in the Russian cybersecurity market, or the prospect of AI replacing infosec experts, you are welcome at the business session of the forum.

The event evolves, but the conditions of participation remain the same. All presentations at PHDays 12 will be in the form of a talk (50 minutes) or a fast track (15 minutes).

## How to apply

To take part in the forum, fill out an application on the CFP page with information about yourself and your proposed talk.

If your research has already been published, specify the relevant conference, journal, or website. If you know of similar work by other researchers, give details and indicate how your approach is different.

You may submit as many proposals as you want.

Speakers delivering 50-minute talks who are not in Moscow will be provided with a transfer to and from the PHDays venue, hotel accommodation, and two additional tickets to the forum.

Don't hesitate to apply, and mark May 19 and 20 on your calendars!

See you at PHDays!

Positive Hack Days 12 will be held in Gorky Park on May 19–20

VK has launched a bug bounty program on The Standoff 365 Bug Bounty, a platform developed by Positive Technologies for uncovering vulnerabilities. A bug bounty program offers rewards to outside researchers for discovering and reporting security flaws before attackers can find and exploit them.

The VK bug bounty consists of over 40 projects. Rewards range from $100 to $30,000, depending on the severity level of reported vulnerabilities.

VK was one of the first companies in Russia to offer rewards to external researchers for reporting vulnerabilities. Since 2013, the company has received over 15,000 bug bounty reports and used them to fix vulnerabilities and strengthen the protection of users' personal data. In total, the company has paid out more than $3 million in bug bounties.

The security of our customers and their trust in VK services have always been a priority for us. We use various tools to strengthen our cybersecurity defenses, and we have long taken advantage of bug bounty programs to test the quality of our services. This method has a real practical benefit. In the last six months, the number of cyberattacks in Russia has increased manifold, and we are happy to see Russian companies launching their own bug bounty platforms. By introducing our bug bounty program on The Standoff 365, we will boost VK's ability to identify vulnerabilities and fix them as quickly as possible.

It takes courage for a company to bring in outside cybersecurity experts. But that's exactly the right way for companies to get a reliable and objective assessment of their security, so they can eliminate vulnerabilities in their IT infrastructure before the bad guys find them and cause irreparable damage. Launching a bug bounty program means thinking in the long term. It's a sign that a company is open and cares about the security of its customers' personal data. The mere fact that you have a bounty program makes people trust you more. We are happy to welcome VK on our platform, and we look forward to continued cybersecurity cooperation with such an experienced bug bounty organizer.

Positive Technologies announced the launch of [The Standoff 365 Bug Bounty platform ](https://bugbounty.standoff365.com/en-US)at PHDays 11 in May this year. Security researchers working on the platform will be rewarded for identifying vulnerabilities and also for demonstrating ways in which they can be exploited. 1600 researchers have already registered for The Standoff 365 Bug Bounty and submitted 49 vulnerability reports, with the first report being sent 60 just minutes after the platform was launched. Positive Technologies and Russian supermarket chain Azbuka Vkusa were the first companies to host their bug bounty programs on the platform.

The creators of The Standoff 365 Bug Bounty expect that 10 to 20 organizations will add their bug bounties to the platform in 2022. By 2025, that number may exceed 100.

Russian social network VK joins The Standoff 365 Bug Bounty


Positive Hack Days 11 will begin in a matter of weeks. This international forum on practical security will be held on May 18–19 in Moscow. The red and blue teams for The Standoff have already been formed, and we are putting the finishing touches to the [cyberrange](https://range.standoff365.com/) infrastructure and the conference program.

As per tradition, PHDays will have three big tracks dedicated to countering attacks (defensive), protection through attack (offensive), and the impact of cybersecurity on business. It is our pleasure to present the first talks.

## How to detect 95% of attacks covering 5% of threat actors' techniques

Oleg Skulkin, Head of Digital Forensics and Incident Response Team, Group-IB, will analyze a short list of techniques (used by almost all threat actors, no matter their sophistication) based on real-world attack scenarios. This provides detection opportunities even if there is very little data.

## IoC scoring

When dealing with indicators of compromise, analysts need to quickly understand the danger posed by the object in question. For this purpose, a special threat intelligence score is used. How exactly the vendor calculates it is often a commercial secret. Nikolay Arefiev, co-founder of RST Cloud, will explain how scoring works using the example of open indicators.

## If you have bootkits

When a computer is infected with viruses at the user level, you can use known methods of counteraction that rely on the kernel API. And what if the OS kernel itself or the firmware is compromised? Anton Belousov, Senior Specialist at Malware Detection, Positive Technologies, will talk about potential vectors of infecting BIOS- and UEFI-based systems with bootkits, and explain how to use the Xen–LibVMI–Drakvuf bundle to monitor malware behavior and what events or signs are indicative of an attempt to introduce a bootkit.

## Sanitizing the Linux kernel

In his report, independent information security researcher Andrey Konovalov will focus on KASAN implementation and practical usage, but will also briefly cover other sanitizers—the main tools for detecting bugs in the Linux kernel. KASAN detects memory safety issues: out-of-bounds and use-after-free bugs in slab, page_alloc, vmalloc, stack, and global memory.

## Open-source intelligence

Andrey Masalovich, CEO, Inforus, will cover 20 practical OSINT techniques leveraging the opportunities of the digital age: image search using neural networks, collecting information from the dark web, detecting cloud storage leaks, tracking a user's digital footprint.

## Qualcomm BootROM

Vulnerabilities in BootROM, the most important component of hardware and software security, can lead to attackers gaining full control over the device. Independent researcher Dmitry Artamonov will discuss the role of BootROM in the Android smartphone boot chain. He will also talk about BootROM vulnerabilities of various mobile device vendors, share his experience of getting access to the JTAG interface in a Qualcomm smartphone, explain how to use it to extract the BootROM image from a modern device, and demonstrate successful exploitation of a one-day vulnerability in BootROM.

## Phishing on official websites

It is generally believed that fake websites are used for phishing. And if the site is genuine? What specific issues can lead to such consequences as website hacking? Independent information security researcher Aleksandr Kolchanov will give a number of examples of not just small companies easily falling victim to such phishing attacks, but large banks and airlines too. He will talk about common and lesser-known problems, including subdomain takeover and attacks on administrators of external services and on URL shorteners.

The co-organizer of PHDays 11 and The Standoff cyberbattle is the [Innostage Group](https://innostage-group.ru/). The business partners of the forum are [Security Vision](https://www.securityvision.ru/), a developer of cybersecurity solutions, [Rostelecom-Solar](https://rt-solar.ru/), a national provider of information security services and technologies, and [MONT](https://www.mont.ru/ru-ru), a distributor of software for any business. The technological partner is [Azbuka Vkusa](https://av.ru/). The partners of PHDays 11 are [Axoft](https://axoftglobal.ru/), [Fortis](https://fortis.ru/), [ICL System Technologies](https://icl.ru/eng/), [InfoWatch](https://www.infowatch.ru/), [Marvel Distribution](https://www.marvel.ru/en/), [R-Vision](https://rvision.ru/), [Gazinformservice](https://www.gaz-is.ru/), [Pangeo Radar](https://pangeoradar.ru/), [Jet Infosystems](https://jetinfosystems.com/?from=ru), Liberum Veritas, [IBS Platformix](https://platformix.ru/), and [USSC](https://www.ussc.ru/).

Stay tuned for more updates!

PHDays 11 talks: bootkit infection, sanitizers for the Linux kernel, the new face of OSINT, and phishing on official websites

Collectors no longer store works of art in safes—galleries and museums are now no less secure than the most impenetrable vaults. But can criminals steal a painting without leaving home?

PHDays 2022 in May will see a revamped edition of The Standoff Digital Art: intrepid researchers will again try to steal cryptomasterpieces right there in the metaverse, while forum guests will be able to visit London without leaving Moscow.

Last year can be safely called the year of NFT. Trade on the NFT art market [reached](https://markets.businessinsider.com/news/currencies/nft-market-41-billion-nearing-fine-art-market-size-2022-1) $41 billion, coming close to the traditional art market. Tokens allow everyone to become owners of unique items. When artists, especially during lockdown, had new opportunities to sell their works, cybercriminals came knocking: phishing and exploitation of smart contract vulnerabilities are now all too common in the NTF sphere.Experts do not rule out the emergence of new methods of fraud in this market.

Most smart contract vulnerabilities are related to the generation of a new collection. At this stage, each item in the collection is assigned a set of characteristics. Having learned to predict them, an attacker can intercept the rarest and most expensive NFTs outside the pricing rules.At PHDays 2022, participants will have to solve even more complex and interesting tasks than last year, and our competition will reach a cosmic level: together with Arcona, we will place digital artworks in augmented reality.

Paintings by Russian artists will appear in the virtual art galleries of Arcona XR Metaverse one week before the forum. The application will let you visit the digital exhibition from anywhere in the world.

Arcona XR Metaverse is a global augmented reality (AR) metaverse. It is based on a unique technology platform that automatically generates an AR layer anywhere in the world. On this layer, on digital land, different interactive projects can run simultaneously: games, historical reconstructions, tourist attractions. Such a large-scale ecosystem increases the demand for AR technology, making it mainstream. Armed with a digitized plot of land and a set of 3D models, the general user can place virtual content anywhere in the world in a few minutes, independently and remotely. All digital plots of land and content in the metaverse are designed as NFT assets, which makes the platform open for integration and partnership with any decentralized platforms.

Blockchain remembers everything. This is what hackers can exploit. The contestants will have to conduct their own investigation and find information that will allow them to take possession of digital paintings.

The forum will also host a unique event: a thematic AR project launched in London remotely from Moscow! An interactive installation generated in a Moscow studio will unfold in front of the audience in the British capital.

Want to see it with your own eyes? Welcome to PHDays 2022!

PHDays 2022: cyberart is at risk. Again

New Year has arrived, which means that the cyberindustry's most eagerly awaited event is already on the horizon: the Positive Hack Days international practical cybersecurity forum, held annually since 2011. This year's event will take place on May 18–19 at the now traditional venue of the World Trade Center Moscow. For those unable to attend in person, an online broadcast of all talks and presentations will be available.

PHDays is a meeting place for hackers, security researchers, key experts, opinion leaders, business heads, government officials, scientists, journalists, and more. Here we not only discuss the latest tech challenges and cybersecurity research, but unwrap the most pressing issues facing business and government, and ways to solve them through information security. Past headliners include: legendary cryptographer Bruce Schneier, co-developer of the FaceDancer tool for USB emulation Travis Goodspeed, co-creator of the concept of public key cryptography Whitfield Diffie, and Embedded Security Lead at NVIDIA Alexander Matrosov.

Last year's anniversary event brought together more than 2,500 guests, and over 35,000 people tuned in to the forum and The Standoff cyberbattle online.

The PHDays formula is time-tested and unchanged: a vast testing ground for experiment, unique insight and expertise, professional on-topic conversations, informal communication with hackers, and lots and lots of practical drills—a place where eyes sparkle with the thrill of discovery.

As in previous years, the eleventh PHDays conference will provide its own surprises, united by the general theme of this year's forum, which we will announce in the near future.

Participants can look forward to a rich program of contests, developed by leading cybersecurity experts. The contests are an important part of the event, helping to visualize the infosec threats around us.

Our traditional mix of open-to-all contests includes: legally attacking a payment system, hacking a smart contract, learning about machine-learning techniques in game-based CTF services, and testing them for robustness. All you need is a laptop, curiosity, and passion.

Observe attacks by ethical hackers on a virtual city-state at The [Standoff](https://standoff365.com/), cyberrange, where the world's largest open cyberbattle will once again unfold. In November, 65,000 people worldwide followed the cyberspace clash between attackers and defenders. Ten powerful teams of white-hat hackers spent 35 hours non-stop testing the robustness of the city-state's systems. This year's confrontation will feature all-new tasks and targets, corresponding fully to the realities of the current threat climate and corporate infrastructure.

Forum participants will earn recognition from colleagues and make useful contacts in an informal setting—the reports and presentations will be seen by thousands of infosec stars from around the globe. Anything goes at the forum: if you have a cybersecurity report on a burning topic that you want to share, just submit an application. Anyone can be a speaker: from budding specialists to established pros. You have until March 1 to [apply](https://cfp.phdays.com/login). The program committee will consider all submissions by April 1.

Watch this space for more details of the new PHDays forum and The Standoff cyberdrills. In the meantime, get your applications in and mark May 18–19, 2022, in your calendars.

See you at PHDays 2022!

Positive Hack Days 2022 to take place on May 18–19

This fall, a major cyberbattle will erupt at Moscow's VDNKh Exhibition Center. If the black-hat hackers win the two-day clash, the entire state of F will be plunged into chaos. The bravest companies in industry, retail, and finance will test the robustness of their systems. Our task is to protect the virtual state and learn how to thwart attacks in real life. The world's largest cybertraining, featuring an expanded business and technical program, will be held in hybrid format.

Now The [Standoff](https://standoff365.com/) is more than a virtual city—it is a whole country, where we will simulate the technological and business processes of real companies from steel and chemical industries, energy, transport and logistics, and municipal services. There, organizations can test their in-house systems: analyze the level of security, trace typical cyberattack chains, correctly verify unacceptable events and their consequences, and assess potential damage.

In 2020, held separately from the Positive Hack Days Forum, the world's largest open cyberbattle moved up a gear. The event was watched live on air by 20,000 captivated viewers.

The defending and attacking teams comprise experts from real companies and independent security researchers from across the globe. The Standoff will present an opportunity for them to hone their skills and, in just a few days, gain unique cyberconfrontation experience. In ordinary life, this would take years.

The Standoff is not only a thrilling and dynamic cyberbattle, but also a platform for dialog about information security. It is here that world-renowned experts, including professionals from Cyber R&D Lab, Hack In The Box, Positive Technologies, and other international companies, pool their vast knowhow. We consider and debate how the cybersecurity industry will evolve. And if that wasn't exciting enough, we also hold contests for participants.

The dramatic rise in the number of cyberthreats in Russia and worldwide requires new security solutions. Together with leading infosec experts, we will take a no-nonsense, no-preaching look at the latest cybersecurity challenges and protection methods. We will talk about pressing business problems, develop the important dialog with the government on import substitution, discuss prospects for investing in cybersecurity, and explain how rising cybercrime affects the investment appeal of the industry.

See you at The Standoff!

Testing the security of the virtual state: The Standoff returns on November 9–10, 2021

We keep elaborating on the topic of AI security and risks and, for PHDays (phdays.com), we put together a track with talks and organized a CTF competition for cybersecurity experts, which addresses the risks of AI. In this article, we will tell you about the competition: what tasks were there and how everything went.

AI CTF has taken place before, and we already published the last description of the format and t[asks on Habr.](https://habr.com/ru/companies/pt/articles/454206/)

## Writeups of tasks

The topics of information security and artificial intelligence are quite popular separately. But when they are combined, adversarial attacks most often spring to mind, although they are not so common in real life. We have tried to highlight other possible risks.

This year's tasks were evaluated dynamically: at the very beginning, all tasks cost the same, and later their cost decreased if they were solved by many players. This saved us from incorrect evaluation of tasks and, in our opinion, made the evaluation fairer.

## 3v1l_k3yb04rd

This task became the most popular among the solved ones but at the same time raised many questions.

The idea of the task is a reference to data privacy and language models that can give away the private information on which they were trained.

According to the description of the task, the players got a model of autocompletion of [words](https://colab.research.google.com/drive/1iZph5io0IO1fYLfZCMGPdYibbwwl0e6A?usp=sharing#scrollTo=RSV1JFFf6nrt), extracted from the keyboard of one of the top managers of some company and the last words he typed. The players were asked to get the password, which would be a flag for getting points.

The task could be solved simply if, in the API to the model response, the players changed the top-n predictions that were displayed to the user (by default, one prediction was displayed). Starting with the phrase that was in the description, it would be possible to obtain several options and use them for the next prediction. Trying different things, at some point, the model would suggest a phrase supplemented with a flag.

Starting with `Somewhere in`, it was possible to get the supplementary options and, going through them, it would be possible to obtain the full phrase. Ideally, the necessary sentence was: `Somewhere in Moscow City hacker Bob broke the system with password`, and the flag would be added to it.

The flag was the phrase `pr1v473_d474_5h0uld_b3_pr1v473`, which should have been enclosed in the tag `AICTF {__}`.

To make the mechanics of the task clear, we decided that it would be cool if the phrase itself was clear: who did what and where. According to this scheme, it is clear where to go. However, we were not able to clearly reflect this scheme in the task description, and the players found some difficulties in solving it: for example, they received only part of a phrase or a phrase in a similar flag format. But still, the task turned out to be solvable for most of the players.

## Game of Cats

In our opinion, this is a fun task for automation, but not everyone managed to complete it.

At the heart of the task is a game, and it is not known how many levels there are in it, and if you start playing manually, you will have to choose a photo at each level, which shows a cat, and click Next. At some point, instead of one image, you get two, three, and up to fifty. At the same time, there is a time limit, and the progress is reset when an error occurs.

We tried to adequately complicate the game so that it wasn't played manually, but that was what some desperate players did.

We took an open [dataset](https://www.kaggle.com/c/dogs-vs-cats), hoping that the players would find a trained model and simply automate the playthrough.

But in fact, there was another hack: they could see that the state of the game did not change, that is, even without using the model, they could just go through the options for each level, save the successful ones, and start again in case of an error.

## AlexNetv2.0

In my opinion, this task should have been the easiest, but we did a dynamic assessment—and we were not mistaken.

The players were given a model and told that it was some new version. The task was from a stegano category—a category of tasks aimed at extracting information.

Having taken a little look at the model, the players could see that the architecture was a regular AlexNet model:

Where is the flag hidden here? Having given it a little thought (and this is the essence of each task), the players can come to the conclusion that the layers in the convolutional network are also images and try to visualize them. There are a lot of functions for visualizing layers, they can be googled (analyticsvidhya.com/blog/2020/11/tutorial-how-to-visualize-feature-maps-directly-from-cnn-layers), and many work out of the box.

Without going far into the deep part of the network, the players could find the flag:

## h4ck3r

We could not leave the classic CTF players without a task, so we thought about this task for a long time.

A service was given that accepted an image as input, but also returned an image with the applied filter and hacker score as output. And only a real hacker could get the maximum score!

What's the first thing a data scientist will think? It is necessary to craft images and get the maximum score—the adversarial attack! But it is difficult to craft images in the case of a black-box model. We tried to hint that this should not be done.

The task category was marked as web and stegano, so a real CTF player would start looking for a bug in the web! Having exploited Path Traversal, which was prudently embedded in the service, the players could obtain the service source code and see how this hacker score was given to the image.

And in addition, there was another interesting detail: the flag was mixed with the picture, but the number of flag characters was set depending on the resulting score.

Possible options:

On the one hand, the participants had to implement the reverse mixing algorithm to get the information from the image.

On the other hand, there was no way to obtain the whole flag, other than to send a few pictures to get the different parts and, subsequently combining them, get the correct full string.

We thank the RuCTF team, who allowed us to take their cool service as the basis for the game!

## BigLittleData

All that was given to the participants was a CSV file with a set of data. What to do?
I see the data—I open it in pandas!

It looked something like this:

And if you are already an experienced data scientist, then most likely you will want to visualize the data. Actually, choosing a visualization method was the right way to solve the task. But everyone understood the task differently, and sometimes very interesting visualization options came out. In our opinion, they could be almost called art.

We agree here that there were a lot of visualization options, but we hoped that the players would choose the most popular ones and just give it a try. This task was also solved, but it was not so easy.

The players could get a flag by tampering with the PCA algorithm.

## Tensorch

In preparing this task, we were inspired by the existing CVEs in frameworks for deep learning. But to simplify the task, we decided not to use real frameworks with a huge code base, but to write small service of our own.

The players were given a binary file that read and ran the user's neural network. They had to find a vulnerability in it and remotely exploit it.

There were two (intentionally embedded) vulnerabilities in the binary file:

The size of the user input was not checked.
When selecting the activation function for the layer, the function index was checked incorrectly.
The first vulnerability allowed us to obtain the addresses of the stack and load the binary file itself in the memory, which in the future would help us to bypass ALSR. The second vulnerability made it possible to jump to an arbitrary address and start executing our own code.

How to exploit the first bug:

1. The structure that stores the neural network lies on the stack. There are also coefficients for the layers.
2.  We will enter the values of our input in a special way and specify a fairly large size so that we get val_from_stack * 1.0 as the network output.
3.  Next, we translate the double sent to us into bytes and thus get the addresses.

How to exploit the second bug:

1. We check the signed int for the fact that it is less than three, but if we enter -1, then the value from the buffer that stores the output of our network will be used as the address of the function.
2. We create a new network with eight outputs and adjust the coefficients so that all the outputs are equal to the address we want to jump to.
3. Then we create a new network with the activation function equal to -1.

Now we can jump to any address, but this is not enough. At the same time, there is one interesting place in the binary file:

 text:00000000000019C1                 mov     edx, 240h   ; nbytes
.text:00000000000019C6                 mov     rsi, rax        ; buf
.text:00000000000019C9                 mov     edi, 0          ; fd
.text:00000000000019CE                 call    _read 
At the time of calling the activation function, the RAX register contains the value that should be passed to the activation function:

.text:00000000000014AD               movq    xmm0, rax       ; a
.text:00000000000014B2                call       rdx                    ; activation function call
Thus, given that in step 1 we learned the address of the stack, we can simply read our ROP chain into the stack. The address must be calculated in such a way that that we overwrite the output address of the _read function. After that, there is another small exercise to leak the address of the libc and read the second chain.

## Results

This year we tried to make the tasks a little more familiar to data scientists and believe that it turned out quite fine. We had three tasks specific to data scientists, one task for automation, and two tasks more familiar to the classic cybersecurity specialists (although in the specified subject).

Within a little more than one day, 19 players managed to report at least one flag. All the tasks were solved, but no one solved everything. On the one hand, there was a risk that someone would solve everything and the game would become boring, but on the other hand, it seems that we managed to make the tasks quite solvable. But next time we will have to come up with a more complex task to definitely make the game a real challenge.

## Prize-winning places were taken by:

1. Жмыхолет Апач
2. pomo_or_not_pomo
3. konodyuk

We awarded the guys with prizes relevant to the competition: [AWS](https://www.amazon.com/AWS-DeepRacer-Evo-Car-Sensor/dp/B081GZSJVL) DeepRacer for the first place, [JetRacer](https://www.amazon.com/JetRacer-Accessories-Following-DonkeyCar-XYGStudy/dp/B0838B2YZQ) AI racing robot for the second, and [Jetson](https://www.amazon.com/NVIDIA-945-82771-0000-000-Jetson-TX2-Development/dp/B06XPFH939) TX2 Module for the third place.

We hope that we help data scientists and cybersecurity experts to become acquainted with each other's professional world and enjoy the game!