Sign in

Tal Skverer

@taltechtreks.com
3.6K followers 102 following 386 posts

Security Researcher | Blog writer | Hacker | DEFCON speaker | Gamer | Silly stuff lover. taltechtreks.com

PostsRepliesMedia
Tal Skverer @taltechtreks.com · 27/04/2026
Look what just landed in GitHub's changelog 👀 A well-needed change to how GitHub Actions constructs the sub claim in its OIDC tokens, making OIDC trust relationships more secure for organizations by embedding immutable identifiers in the claim.
111
Tal Skverer @taltechtreks.com · 29/11/2025
It happened! My #DEFCON 33 talk on hacking Pokémon Go is live! Looking back, it was an absolute pleasure to dip into that 2016 nostalgia and share with the security community the untold story of Team Unknown6, on stage in front of a packed full audience. m.youtube.com/watch?v=2En9...
m.youtube.com
DEF CON 33 - Breakin 'Em All – Overcoming Pokemon Go's Anti Cheat Mechanism - Tal Skverer
YouTube video by DEFCONConference
020
Tal Skverer @taltechtreks.com · 28/10/2025
וואו, איזה אווירה ואיזה קהל ב-Reversim Summit היום! שמחתי מאוד על הזכות להיות פה ולהציג בפניכם את המחקר שלי, תודה!
120
Tal Skverer @taltechtreks.com · 27/10/2025
איזה כיף שהגיע הכנס השנתי של Reversim Summit! כנס שתמיד נהדר להיות בו עם הרצאות מעניינות ומעולות, והפעם הוא מתוק במיוחד כי יוצא לי להרצות בו בפעם הראשונה! מוזמנים לבוא לשמוע על איך נראה האקינג למשחקי אנדרואיד, ואיך ביליתי את ימי הקיץ אי שם ב2016 כדי לפרוץ לפוקימון גו. מחר, 27/10, ב13:20 באולם הגדול!
110
Tal Skverer @taltechtreks.com · 24/10/2025
Had a great time presenting the OWASP Top 10 NHI Project today at #LASCON Austin with Tomer Yahalom! We've had the pleasure to give the talk to a great crowd and amazing people. Looking forward to the second day filled with interesting talks!
010
Tal Skverer @taltechtreks.com · 16/10/2025
The #MCP framework took the world by storm. But is it safe? We downloaded over 5 thousand of the highest starred MCP implementations to answer that question! astrix.security/learn/blog/s...
astrix.security
State of MCP Server Security 2025: Research Report | Astrix
5K+ MCP servers analysis: 53% use insecure hard-coded credentials. Read the full 2025 research and download the open-source MCP Secret Wrapper to mitigate risks.
100
Tal Skverer @taltechtreks.com · 20/08/2025
I am absolutely delighted to share that I'll be giving a talk at Reversim Summit 2025 on breaking Pokémon Go's anti-cheating mechanism! On a personal note, every time I visit Reversim, it shows me just how awesome, welcoming, and genuinely interesting community-driven events can be.
110
Tal Skverer @taltechtreks.com · 09/08/2025
Breaking 'Em All! The blog posts going deeper into hacking Pokemon Go that accompany my #DEFCON talk are live on my blog! taltechtreks.com/2024/04/06/H... taltechtreks.com/2025/08/09/H... Happy to get questions on the research!
taltechtreks.com
Breaking Pokémon Go Anti-Cheating System - Part 2
Overcoming Niantic's Anti Cheating Mechanism
061
Tal Skverer @taltechtreks.com · 09/08/2025
Talking in an hour at #DEFCON about Pokemon Go on track 3. Couldn't be more excited! I've worked on it multiple nights in the past months, and I hope you'll enjoy it! If you're not here in person, you can catch the talk live at www.youtube.com/live/fzbrrKP... or www.twitch.tv/defcon_dctv_...
twitch.tv
DEFCON_DCTV_Three - Twitch
DEFCON TV Channel 3: Track 3 Talks
051
Tal Skverer @taltechtreks.com · 09/08/2025
Giving a talk on Pokemon Go, I just had to convert myself to a Pokemon trainer. The amazing Ryan Rockenbaugh surprised me with a limited edition sticker of Trainer Tal Sticker wall at #DEFCON got 2, and I will be giving what's left tomorrow at my talk Breaking 'Em All! 11:30 Track 3. See you!
060
Tal Skverer @taltechtreks.com · 07/08/2025
Excited and humbled to speak once again in DEF CON. Talk is on Saturday at 11:30, track 3. I'll be in the conf till Sunday, hit me up if you want to chat over the talk or any other project.
020
Tal Skverer @taltechtreks.com · 22/07/2025
The #DEFCON 33 schedule is live and I'm excited to announce I'll be giving a talk this year on unique research I was a part of a few years back! "Breakin 'Em All – Overcoming Pokémon Go's Anti-Cheat Mechanism" Join me on stage - Sat, 11:30 AM, Track 3 defcon.org/html/defcon-...
011
Tal Skverer @taltechtreks.com · 11/06/2025
Looks like I'll be at Hacker Summer Camp this year! Exciting #DEFCON33 ☺
050
Tal Skverer @taltechtreks.com · 06/06/2025
Just got back from #AppSecIL2025! Ended up 4th place in the #CTF 🎉 Solved 12/15 challenges alone - Android pwn, JS sandbox escapes, cache poisoning, XSS bypasses. The usual suspects: SQLi, LDAP injection, XXE. Had a blast! Looking forward to the next one. Writeup: taltechtreks.com/2025/06/04/a...
taltechtreks.com
AppSec IL 2025 CTF - Writeup
A writeup on all challenges I solved in the 2025 OWASP CTF
021
Tal Skverer @taltechtreks.com · 20/05/2025
I'm building something cute and new, but as opposed to my normal ways, this time I'm using #Cursor heavily and the flow blows my mind. Although I'm certain that without prior good knowledge of coding it wouldn't work that well.
000
Tal Skverer @taltechtreks.com · 02/05/2025
Ah, I wondered when this was due to come out. Can't wait to experiment! www.anthropic.com/news/integra...
anthropic.com
Claude can now connect to your world
Today we're announcing Integrations, a new way to connect your apps and tools to Claude. We're also expanding Claude's Research capabilities with an advanced mode that searches the web, your Google Wo...
000
Tal Skverer @taltechtreks.com · 01/05/2025
Excited to present my research tomorrow at #RSAC! Come by to hear John and me share conclusions and insights on the first publicly available report on Non-Human identity security! See you there, May 1, 10:50 am at Moscone West 3004!
030
Tal Skverer @taltechtreks.com · 23/04/2025
For today, a bit Friends content! Or, uhh, actually, computer science! What’s the largest sofa you can pivot around a corner? Heard about this problem in the past, but I thought it's one of those we will never be able to prove. (At least until quantum computers arrive to solve some NP problems)
quantamagazine.org
The Largest Sofa You Can Move Around a Corner | Quanta Magazine
A new proof reveals the answer to the decades-old “moving sofa” problem. It highlights how even the simplest optimization problems can have counterintuitive answers.
010
Tal Skverer @taltechtreks.com · 17/04/2025
Took a bit of hiatus from posting here, was quite busy the past few months (CFP season amiright…) But I do have more cool stuff ready for sharing as well as writing 2 blog posts about projects I worked on lately. Looking forward to sharing with you all
000
Tal Skverer @taltechtreks.com · 14/03/2025
Great time presenting OWASP NHI Top 10 at #SnowFROC! If you're here, come say hi!
000
Tal Skverer @taltechtreks.com · 08/03/2025
Well, that was an hour of my life well spent. An amazing adaptation of minesweeper to include old-school RPG elements. UPDATE: Had to go back and play, another 4 hours to achieve a perfect clear. Incredible game!
danielben.itch.io
Dragonsweeper by Daniel Benmergui
A roguelike minesweeper adventure
010
Tal Skverer @taltechtreks.com · 06/03/2025
Great thread on the emerging new details on the ByBit breach
000
Tal Skverer @taltechtreks.com · 28/02/2025
A cute post showcasing the basis of every website - HTML! The site goes over (almost) all HTML tags, by using them. Seeing this, it’s unfortunate that text inputs in websites don't allow most tags but rather encapsulate how the final result is displayed for users. Give us more, please!
iamwillwang.com
Every HTML Element
010
Tal Skverer @taltechtreks.com · 27/02/2025
Managed to squeeze in 2 talk submissions to #fwd:cloudsec just before the first round CFP closes. Hoping for good news, will be the perfect reason to finally attend in person!
000
Tal Skverer @taltechtreks.com · 24/02/2025
Check out this wild project: #Steam #Brick. The author transformed the Steam Deck to a brick that still connects to screens or VR while reducing size by a third Projects like this (author used steam-provided data) are a great example of "right-to-own", critical to enhance electronics sustainability
crastinator-pro.github.io
No screen, no controller, and absolutely no sense, just a power button and a USB port.
Brick your Steam Deck - one way or another…
010
Tal Skverer @taltechtreks.com · 21/02/2025
#OAuth #phishing apps are coming to get you - now on #X! It's annoying to see yet another case of how easy it is to fake consent screens that look identical to real companies. Why we still lack mandatory verification processes for new OAuth apps is beyond me. x.com/thealexbanks...
020
Tal Skverer @taltechtreks.com · 20/02/2025
#OIDC identity providers in #AWS have been getting more attention lately and rightfully so: It’s unfortunately surprisingly easy (and common) to create misconfiguration within the necessary conditions on OIDC-based #IAM roles trust policy.
offensai.com
RogueOIDC: AWS Persistence and Evasion through attacker-controlled OIDC Identity Provider
This research shows what an attacker can achieve after creating a malicious OIDC identity provider in AWS and how they can do it. The article presents novel techniques and tools for persistence and ev...
110
Tal Skverer @taltechtreks.com · 17/02/2025
So, #OAuth implementations seem to always have some edge case or quirk making it vulnerable in odd ways. This particular exploit leveraged an open redirect in an app's OAuth callback, combined with #Google OAuth's quirks to steal authorization codes on visiting a malicious website.
blog.voorivex.team
OAuth Non-Happy Path to ATO
Learn how small errors in OAuth implementation can lead to significant security vulnerabilities like one-click account takeover in web applications
074
Tal Skverer @taltechtreks.com · 15/02/2025
#DeepSeek roasts me based on my #UserAgent: Your browser history is so bare even your referrer ghosted you, your 2560x1440 screen is just compensating for how bland Firefox 134 on Windows 10 is, and 16 CPU cores in Tel Aviv? Congrats, you’re the NPC overclocking spreadsheets in the Silicon Wadi.
jasonthorsness.com
DeepSeek My User Agent
DeepSeek My User Agent
021
Tal Skverer @taltechtreks.com · 14/02/2025
Cute bit of #internet history, this post shows why #Git gives you 0.1 seconds to decide if it runs a fixed command. Turns out, backwards compatibility in code is hard, especially if the code is distributed widely, and can lead to these kind of funny instances (and other that can cause real damage)
blog.gitbutler.com
Why is Git Autocorrect too fast for Formula One drivers?
Why does Git's autocorrect wait 0.1s before executing a mistyped command? Let's dig in.
000
Tal Skverer @taltechtreks.com · 12/02/2025
I love #hacking #embedded systems, and also #gaming when time permits. Imagine my delight seeing a post that blends both! Two decades after its release, someone revisits the idea of “just” using the original #Xbox CPU's #JTAG connector to debug it.
blog.ret2.io
JTAG 'Hacking' the Original Xbox in 2023
Released in November 2001, the original Xbox was Microsoft’s first venture into the game console industry. With its hardware closely resembling a cheap but v...
140
Tal Skverer @taltechtreks.com · 09/02/2025
The Dead Internet Theory is just a meme, right? Well…something strange happens over at PhysicsForums. Once an active forum for physics students, now a quieter place. A year ago, posts appeared, made by users long after their last login. LLM-generated replies back-inserted into the DB Must read!
hallofdreams.org
PhysicsForums and the Dead Internet Theory
An exposé no one will read, about the widespread falsification of user posts in PhysicsForums, a scientific community founded in 2001. This is a microcosm of the death of the human-written Internet.
020
Tal Skverer @taltechtreks.com · 08/02/2025
Still surprised with how people use LLMs for cool stuff. This one uses o1 and makes it play Codenames with itself. If you played it, you know it's tricky to give good hints, and o1 crushed it. Wonder how it works when not playing itself. If I played someone with my brain, the game would be easy :)
suveenellawela.com
Codenames AI - OpenAI's o1 Playing Codenames
I got OpenAI's o1 to play codenames and the results were surprisingly good
011
Tal Skverer @taltechtreks.com · 07/02/2025
The #GPT #o1 model was very exciting initially, but I completely stopped using it. The long wait and the laziness of my prompts made the chat-based models much better. This post explains how to interact with o1. In short - it comes down to good #prompts so o1 one-shots the response.
latent.space
o1 isn’t a chat model (and that’s the point)
How Ben Hylak turned from ol pro skeptic to fan by overcoming his skill issue.
000
Tal Skverer @taltechtreks.com · 04/02/2025
Honey, wake up - #clickjacking is back. A decade ago, #clickjacking was the hot new #attack—tricking users into clicking invisible #frames layered over legit websites, unknowingly performing actions on their social media, corporate accounts, or even banks. 1/3
paulosyibelo.com
Blog: DoubleClickjacking: A New Era of UI Redressing
data:blog.metaDescription
131
Tal Skverer @taltechtreks.com · 02/02/2025
#Hacker #News today: Well-written piece about a paper on #E2EE (end-to-end encryption) when it comes #AI. AI circled us back to the old problem: devices are too weak to handle computations, so we offload them to more powerful systems—essentially handing most sensitive data over to big data. (1/5)
blog.cryptographyengineering.com
Let’s talk about AI and end-to-end encryption
Recently I came across a fantastic new paper by a group of NYU and Cornell researchers entitled “How to think about end-to-end encryption and AI.” I’m extremely grateful to see th…
130
Tal Skverer @taltechtreks.com · 01/02/2025
#Hacker #News daily: A month with Devin – an interesting log from a team that put one of the most hyped startups to the test. The conclusion? Potential is definitely there, but it’s still wrestling with similar challenges ML algos suffered from in the past - mostly the need to manually verify work.
answer.ai
Thoughts On A Month With Devin – Answer.AI
Our impressions of Devin after giving it 20+ tasks.
000
Reposted by Tal Skverer
Nathan McNulty @nathanmcnulty.com · 24/01/2025
This provides important insights 💡 CA policies cannot block anything until AFTER authentication occurs This means CA cannot help with password spray/credential stuffing. This is why we have Password Protection and Smart Lockout. learn.microsoft.com/... learn.microsoft.com/...
learn.microsoft.com
Password protection in Microsoft Entra ID - Microsoft Entra ID
Learn how to dynamically ban weak passwords from your environment with Microsoft Entra Password Protection
1273
Tal Skverer @taltechtreks.com · 24/01/2025
This old post from 2006 jumped on my feed and I recalled reading it back then, I wasn't too familiar with binary search, but was still shocked a simple #overflow bug was so prevalent. 20 years later, I totally see how generative #AI might cause thousands more of these issues to continuously pop up
research.google
Extra, Extra - Read All About It: Nearly All Binary Searches and Mergesorts are
Posted by Joshua Bloch, Software EngineerI remember vividly Jon Bentley's first Algorithms lecture at CMU, where he asked all of us incoming Ph.D. ...
0188
Tal Skverer @taltechtreks.com · 23/01/2025
Oh boy, Sam Curry comes again with yet another critical #vulnerability, now in Subaru vehicles, basically allowing full remote control over the cars. #Automotive security is no joke, but manufacturer's still live by security standards from the beginning of the century. samcurry.net/hacking-subaru
samcurry.net
Hacking Subaru: Tracking and Controlling Cars via the STARLINK Admin Panel
On November 20, 2024, Shubham Shah and I discovered a security vulnerability in Subaru’s STARLINK admin panel that gave us unrestricted access to all vehicles and customer accounts in the United State...
1155
Reposted by Tal Skverer
OWASP® Foundation @owasp.org · 17/01/2025
Got secrets? An API? Tokens? Or a fancy new AI bot? The OWASP NHI Project team just dropped the Non-Human Identities Top 10 Project, breaking down the biggest risks + how to secure them. :closed_lock_with_key: Check it out here: t.co/wmUD9x9sJO
t.co
OWASP Non-Human Identities Top 10 | OWASP Foundation
The primary goal of the
0135
Tal Skverer @taltechtreks.com · 12/01/2025
#Hacker #News daily: Sometimes, you read amazing things online and just go with: What. This is one of these things. Apparently, it's entirely feasible for build a functional CPU using...regex? This madman has built it, and an entire chess engine on top of it. Simply incredible.
nicholas.carlini.com
Regex Chess: A 2-ply minimax chess engine in 84,688 regular expressions
I wrote a (list of) regular expressions that will play a (not very good) chess game by running a 2-ply minimax search.
010
Tal Skverer @taltechtreks.com · 08/01/2025
Excited to introduce the #OWASP Non-Human Identities Top 10 for 2025! 🚀 What risks do devs face using non-human identities? I tackled this together with other experts to shine light on this emerging issue. Check out the project's page for a sneak peek, or click on top to dive deeper into the risks
owasp.org
OWASP Non-Human Identities Top 10 | OWASP Foundation
The primary goal of the
083
Tal Skverer @taltechtreks.com · 08/01/2025
#Hacker #News daily: these 24 fun facts about SQLite, the most used database engine today. My favorites? #7/8 - It's not easy to contribute code to SQLite. #14 - Reading/writing blobs with SQLite can be faster than `fread` #19 - The reason why the prefix for temp files is `sqlite_` backwards
avi.im
Collection of insane and fun facts about SQLite - blag
Some of the interesting and insane facts I learned about SQLite
020
Tal Skverer @taltechtreks.com · 05/01/2025
Back with some #Hacker #News. This time - some theory! I really like clustering algorithms. They tend to make order into data that can be really confusing. This blog explores a very important theorem from 2002 - showing no clustering algorithm can essentially fulfill all we want from it. (1/3)
blog.codingconfessions.com
The CAP Theorem of Clustering: Why Every Algorithm Must Sacrifice Something
No clustering algorithm is perfect and you must make a trade-off.
230
Tal Skverer @taltechtreks.com · 27/12/2024
Been a while since I've read some papers. My #LLM based monitoring pinged this paper on solutions for distributed cloud. Here's my review. #CompSky Token-based identity management in the distributed cloud A short article proposing a method for #IAM on #distributed #cloud environments. (1/5)
arxiv.org
160
Reposted by Tal Skverer
Martijn Grooten @martijngrooten.bsky.social · 24/12/2024
I love the annual tradition of @lorenzofb.bsky.social @zackwhittaker.bsky.social and @carlypage.bsky.social highlighting the best cybersecurity stories (and, in quite a few cases, thorough investigations) that other people wrote techcrunch.com/2024/12/24/t...
techcrunch.com
These are the cybersecurity stories we were jealous of in 2024 | TechCrunch
The very best work from our friends at competing publications.
02110
Tal Skverer @taltechtreks.com · 25/12/2024
#Hacker #News daily: a directory for hosting verified SSH keys to email mapping. Might currently be a little niche, but I think that a wide adoption of this could ease both user management for SSH servers, and for CLI applications.
keypub.sh
1162
Tal Skverer @taltechtreks.com · 21/12/2024
#Hacker #News Daily: This time, turns out the internet is full of wrong bi-directional BFS implementations! This exceptional blog post covers the issue from the very beginning. Definitely worth reading! zdimension.fr/everyone-get... (1/4)
zdimension.fr
Everyone gets bidirectional BFS wrong
People really need to stop blindly copying code from the Internet.
140
Tal Skverer @taltechtreks.com · 19/12/2024
Today, rspack, a rust-based tool that helps speed up the process of preparing code for websites and apps, has announced a compromise to two of their packages - rspack/core and rspack/cli. github.com/web-infra-de...
github.com
Release v1.1.8 · web-infra-dev/rspack
Security Vulnerability Report Overview This is a re-release version of 1.1.6 On 12/19/2024, 02:01 (UTC), we discovered that our npm packages @rspack/core and @rspack/cli were maliciously attacked. ...
100