Daniel "d3v" Chactoura retweeted
Presence confirmed at BSides BSB 2026!
@kaardeco will demonstrate how a simple diagram file can turn into a real attack vector in Excalidraw!
The talk focuses on dissecting a Stored XSS in manipulated .excalidraw files to execute arbitrary JavaScript.
The presentation covers the entire exploitation path, from file upload to execution inside an iframe, practical uses in phishing, data harvesting, and social engineering.
Beyond the attack, we’ll dive into a complete defensive perspective. Risk mitigation, proper use of sandboxing and secure handling of untrusted content.
100% practical content for anyone living the daily challenges of AppSec, DevSecOps, and Offensive Security.
With extensive background in technology and information security, Allan Kardec currently serves as a Tech Lead for Secure Development, managing DevSecOps pipelines, risk analysis, and threat modeling. Instructor and researcher with hands-on experience in attack simulations and training defensive teams.
See you there!
BSides BSB Schedule: bsidesbsb.com/#agenda
Daniel "d3v" Chactoura retweeted
That moment when the security patch becomes an entry point…
So, Vivaldo Chagas (Nosferatu) and Lucas William found a Stored XSS vulnerability in version 30.2.4 that cleanly bypassed a previous fix from draw.io itself.
draw.io is an open-source diagram editor embedded in half the corporate tools out there: Confluence, Jira, Notion, GitLab, internal wikis.
What made this finding interesting wasn’t simply an unsanitized innerHTML. draw.io actually escapes HTML correctly across most of the codebase.
The real issue was an inconsistency between two parts of the application. One component decides whether a selection can be treated as HTML by looking only at editable cells. Another component, responsible for actually processing the labels, iterates over every cell in the selection, including locked ones with editable=0.
One trusts the wrong assumption made by the other. A classic confused deputy scenario!
In practice, all it takes is two cells in the same diagram: one editable cell to enable HTML mode, and one locked cell carrying the payload.
Simply opening the file is enough. No clicks required. The JavaScript executes in the application’s origin, giving access to whatever is available client-side: cookies, session data, localStorage, and more.
Check out the article on our blog: yokai.hakaisecurity.io/en-dr…
Wanna put this in practice? Take a look in the @hackingclub_io's chall “Drawing”:
app.hackingclub.com/training…
Daniel "d3v" Chactoura retweeted
New Laravel one-day exploit by Yokai Research! 🔥
Full analysis, PoC, and exploit available here: github.com/hakaioffsec/GHSA-…
We analyzed GHSA-f9x3-7vq9-cph6, a critical authentication bypass in Laravel affecting TokenGuard, the native mechanism used to protect auth:api routes.
The issue was originally discovered by KIKOmanasijev(github.com/KIKOmanasijev) and disclosed through GitHub Security Advisory.
After the advisory went public, our team reproduced the vulnerable behavior, analyzed the authentication flow, identified the conditions required for exploitation, and developed a working PoC and exploit.
We break down how the bypass happens, why TokenGuard can fail to enforce authentication as expected, and what that means for applications relying on affected configurations.
No CVE has been assigned at the time of publication.
Authors: @biscoitomesmo and @Gabriel52489026
Blog post:
yokai.hakaisecurity.io/en-gh…
#HakaiSecurity #YokaiResearch #Laravel #SecurityResearch #authbypass #tokenguard
Daniel "d3v" Chactoura retweeted
It’s tomorrow! Three major talks await you at BSides Curitiba!
It’s the eve of BSides Curitiba, and excitement about reconnecting with the cybersecurity community is running high!
Gabriel Santana, aka deneuv, will be on stage sharing a 100% practical, real-case-based analysis of vulnerabilities discovered during Bug Bounty research in Cloud environments. Less theory, more practice, showing how these flaws appear and behave in reality and the valuable details that usually go unnoticed during recon.
@alkarramaxjr, aka Maxin, will show what really happens during an NTLM Relay attack, starting from the basics, analyzing the traffic in Wireshark, breaking down the NTLM handshake, and understanding how the challenge-response mechanism works during authentication.
And we'll also have @KiltzxH, taking the stage to discuss how advanced techniques, such as context building, simulated authority, interaction fragmentation, and indirect induction are being used to bypass guardrails and induce models to perform actions outside their operational guidelines.
Set your alarms, prepare your questions, and come chat with us!
Who has already confirmed their attendance? Comment below so we can meet up at the event!
@BSidesCWB | Tomorrow, 10/17
Check out the schedule: bsidescwb.com.br/programacao…
Daniel "d3v" Chactoura retweeted
Know to Attack: NTLM Relay
What actually happens behind the scenes when an NTLM Relay attack works?
At BSides Curitiba, @Am_IMax will open Wireshark and break down the NTLM authentication flow from the ground up, looking closely at the handshake and the challenge-response mechanism that sits at the core of the protocol.
The talk will then move from protocol analysis to hands-on exploitation, starting with the classic SMB-to-SMB relay scenario and progressing into cross-protocol attacks, showing how authentication can be relayed across different services as part of a broader attack chain.
The goal is to understand how and why the attack works, what happens at the protocol level, and how that knowledge can be applied during real offensive security assessments.
Alkarramax Junior, aka Maxin, is a versatile Red Team operator with four years of experience in offensive security, driven by a strong curiosity to understand how things work under the hood. He explores a broad range of topics, from networking and infrastructure fundamentals to Web application exploitation. Outside cybersecurity, he is also a Souls player, Judo instructor, and Zelda fan, especially The Minish Cap.
Will you be at BSides Curitiba? Don’t miss it!
BSides Curitiba: @BSidesCWB
Schedule: bsidescwb.com.br/programacao…
Daniel "d3v" Chactoura retweeted
The art of deceiving IA!
@KiltzxH will demonstrate how techniques such as context building, simulated authority, fragmentation, and indirect induction can bypass AI guardrails without needing explicit malicious instructions.
If Prompt Injection is just the beginning, social engineering is the next level. He will discuss the limitations of traditional protections and the paradigm shift: AI security as understanding intent and behavior, not just filtering.
Carlos Eduardo, also known as Kiltz, is 25 years old and a pentest analyst at Hakai. Passionate about hacking since childhood, he built his journey exploring vulnerabilities, participating in CTFs, and deepening his knowledge in offensive security with a focus on web/cloud environments.
Will you be at BSides Curitiba? Don't miss it!
BSides Curitiba @BSidesCWB
Check out the schedule bsidescwb.com.br/programacao…
Daniel "d3v" Chactoura retweeted
Bug Bounty in the Clouds!
In this presentation, Gabriel will demonstrate, in a practical way and based on real-world cases, several vulnerabilities he uncovered during his Bug Bounty research in Cloud environments.
The goal is to move beyond theory and show how these flaws manifest in practice, how he discovered them, what was achievable through each vulnerability, and which details often go unnoticed during the reconnaissance process. He will also address how to think beyond the web application and view Cloud infrastructure as an integral part of the
attack surface.
Gabriel Santana, also known as Deneuv, is a pentester and security researcher with experience in Bug Bounty and a focus on Web Hacking and Cloud Security.
If you are at BSides Curitiba, don't miss this technical presentation!
BSides Curitiba @BSidesCWB
Check out the schedule at bsidescwb.com.br/programacao…
Daniel "d3v" Chactoura retweeted
Mirror, mirror on the wall!
Gabriel Alves, an analyst on our team, recently reported two CVEs: CVE-2026-63642 and CVE-2026-63643. Both are Server-Side Request Forgery (SSRF) vulnerabilities identified in MagicMirror².
How do they work?
The flaws are related to URL processing in the newsfeed and calendar modules, allowing requests to be made by the server to attacker-controlled destinations.
What is the impact?
An attacker can use the application to make requests against internal resources that are not directly exposed to the Internet, expanding the attack surface and enabling access to services and information available from the environment where MagicMirror² is running.
Gabriel is passionate about cybersecurity, with a focus on offensive security, pentest, and vulnerability research. Has practical experience in CTF environments, application analysis, and flaw exploitation, always seeking to refine skills
and stay current with new techniques in the field.
Author: @Gabrie0x6c
Check it out: surl.li/ewsaed
Daniel "d3v" Chactoura retweeted
In the last few weeks, we had @gankd_ representing us in Croatia at @BSidesZadar and in Poland at @BSidesKrakow!
He unveiled 9 real, critical-severity security vulnerabilities discovered in widely used libraries, ranging from Remote Code Execution (RCE) via deserialization and template injection to Local File Read, SSRF, and Denial of Service (DoS).
He also showed a 6-step methodology to identify and exploit vulnerabilities in any PDF export feature.
And on top of that, he even won 1st place in the CTF at Security BSides Kraków! 🔥
Wanna know more about this research? Take a look at our blog!
yokai.hakaisecurity.io/explo…
Daniel "d3v" Chactoura retweeted
Vulnerability in OWASP DefectDojo!
Our analysts have found an Improper Privilege Management flaw identified in version 2.59.0, which resulted in the CVE-2026-16764.
The vulnerability allows an authenticated low-privileged user to escalate their permissions to an administrative level by manipulating the is_staff parameter within the UserSerializer function, exploiting insufficient API validation.
As a consequence, an attacker can gain full administrative access to the platform, view and modify vulnerability reports across the entire organization, compromise security triage workflows, and potentially hide or manipulate critical findings.
In this post, we detail the root cause of the vulnerability, its exploitation via API, the fix implemented by the DefectDojo team, and the main recommendations for risk mitigation.
Authors:
Lucas Dantas is an Information Security professional with 2 years of experience at Hakai Security, working in Application Security (AppSec), Secure Development, and Offensive Security. Specialist in Threat Modeling, with experience in Secure SDLC, Code Review, Security Maturity Assessment, Web Pentesting, and vulnerability research
Vivaldo Chagas, pentester at Hakai Security, with a background spanning application security, red team operations, and penetration testing. A native of Amazonas, Brazil, he built his career at the intersection of offensive security and attack surface analysis, relying on hands-on experience in corporate environments.
Check out the technical details and the impact of this flaw in our full article: yokai.hakaisecurity.io/en-cv…
Wanna try out for yourself? Come check the challenge in Hacking Club!
app.hackingclub.com/training…
Daniel "d3v" Chactoura retweeted
About a year ago, Hakai Security introduced the Beerus Framework, an initiative created to centralize security tools and streamline the penetration testing process directly on Android devices.
Now, Beerus expands its capabilities to iOS. Following the same approach as the Android version, the framework aims to gather essential tools into a single environment and simplify device preparation for security testing, reducing complex configurations and operational friction during assessments.
Authors:
Daniel França Lima
Daniel, known as Daniboy, is 19 years old, Brazilian, and works as a Pentester and Researcher at Hakai Security. Passionate about understanding how things work, building tools, and manipulating systems. In his free time, he enjoys gaming, especially titles like Hollow Knight, and he also loves foxes.
Gabriel Rodrigues @Gabriel52489026
Gabriel Rodrigues, researcher and Pentester at Hakai Security. He has one year of experience in offensive security and holds the CPTS certification. As he says himself: “My friends call me Texugo (Badger) because one of them once said that I look like one."
Check out the article on our newly redesigned blog: yokai.hakaisecurity.io/enbee…
Github: github.com/hakaioffsec/beeru…
Daniel "d3v" Chactoura retweeted
Hakai Security research recognized by Apple!
Our analyst Gabriel Rodrigues, from Yokai Research team, was acknowledged by Apple in the security content for macOS Golden Gate 27, following security research related to iCloud that was reported to the company.
The technical details of the vulnerability cannot be disclosed yet, so for now, the finding itself will have to remain under wraps.
But as soon as we are able to disclose, we'll bring you some technical breakdown for sure! 👀
Congratulations, @Gabriel52489026, on the research and the recognition! 🔥
Apple Security Content:
support.apple.com/en-us/1490…
#HakaiSecurity #SecurityResearch #AppleSecurity #CyberSecurity
Daniel "d3v" Chactoura retweeted
Lecture confirmed!
For years, Threat Hunting was based on IoCs. However, indicators only show the tracks of attacks that have already happened.
In this lecture, we will explore how to analyze Threat Actor behavior through TTPs, infrastructure, tools, and operational patterns to identify early-stage campaigns and anticipate attacks before new IoCs even exist.
All demonstrated with real-world cases and hands-on demos using QuimeraX and GOTEL.
Mind the Sec 2026
Track: Fraud | Sub-track: Solutions
Founder of QuimeraX and Crowsec, partner at Hakai Security, and boasting over 13 years of experience in cybersecurity, including hands-on work in Red Team and Threat Intelligence, plus Hall of Fame recognition from Microsoft, Adobe, and the US DoD, Carlos brings the practical perspective of someone who works on the offensive front lines every single day.
Author: Carlos Vieira @carlos_crowsec
SChedule: mindthesec.com.br/agenda-mod…
Mind The Sec 2026 Tickets: mindthesec.com.br/
@MindTheSec
@quimerax_intel
Daniel "d3v" Chactoura retweeted
Confirmed Talk at the 1st Security MeetUP Edition!
The developer is no longer just someone who writes code, but a strategic component in the attack chain.
In this talk, Carlos Vieira (Cadu) will explore how cybercriminals are targeting the dev ecosystem through:
Software supply chain attacks.
Malicious dependencies and repository compromise.
New threats involving AI, MCP (Model Context Protocol), and skill/tool poisoning techniques.
Topic: When the Developer Becomes the Target: Supply Chain, AI, and Poisoning in the Development Pipeline
Who is Cadu?
Co-founder of Grupo G7 (holding company of Hakai Security, QuimeraX, and Crowsec), he has over 11 years of experience in Cybersecurity, Pentesting, and Application Security. He currently works in Threat Intelligence at QuimeraX. A researcher featured in Microsoft's Hall of Fame, he has reported critical vulnerabilities to giants such as Adobe, Ubiquiti UniFi, and the US Department of Defense.
Carlos Vieira: @carlos_crowsec
Hakai Security is also an official supporter of the 1st edition of the Security MeetUP, organized by Avenue.
Date: September 24th, at 7:00 PM
Location: Avenue Office — Rua Maria Carolina, 624, Jardim Paulistano, São Paulo/SP
Format: 100% in-person (Limited spots!) - Registration open until Sept 23rd, at 12:00 PM.
Secure your spot and come exchange knowledge with us:
meetup.com/pt-br/avenue-tech…
Daniel "d3v" Chactoura retweeted
TALK ANNOUNCEMENT!
Meet Lyra: an open-source, cargo-native LLVM IR obfuscator for Rust by @biscoitomesmo.
How it works:
Lyra intercepts each crate during a standard cargo build via RUSTC_WRAPPER, applying a configurable pipeline (string encryption, basic block shuffling, indirect branch obfuscation). It then recompiles with llc + clang and swaps the object before the MSVC linker runs.
Zero Rust compiler modifications.
Zero code changes in your project.
Driven by a single CLI flag.
Speaker: Rafael Félix is a researcher and partner at Hakai Security. Specialized in Rust/C/Assembly offensive tools, Windows internals, EDR evasion, and malware dev (author of Coffee COFF loader, browservoyage, CVE-2024-21338 research). Speaker at DEF CON, BSides LV & Ekoparty.
Sept 19, 2026 at 4 PM (BRT / UTC-3)
Schedule: bsidesjp.com.br/#agenda
Author: Rafael Felix @biscoitomesmo
Daniel "d3v" Chactoura retweeted
Vibe coding brought the barrier to entry down to practically zero. Brilliant. It also hand-waved security responsibilities into non-existence.
Look at applications built on Supabase: the database gets directly exposed through automated APIs via PostgREST.
When the anon role is handed excessive permissions with zero Row-Level Security configured, that public key isn't
just a key—it's a master key. Attackers don't need credentials, elevated privileges, or your source code. They just
need access to the public interface you left wide open.
In Backend-as-a-Service architectures, your database is your first—and final—line of defense. Act like it.
In this talk, Thiago will walk you through the actual attack surface and demonstrate Chupabase (an automated
auditing tool that dissects the bundle, reconstructs the API, and ruthlessly tests endpoints under the anon role). He'll
wrap it up with essential defense blueprints and recommendations you should have implemented from day one.
A bit about Thiago Bispo:
Offensive security professional with solid experience in application security, including secure code analysis, threat
modeling, and architecture security analysis. Possesses strong knowledge in web application, API, and infrastructure
pentesting, as well as Active Directory and cloud security.
Works on integrating security into the software development lifecycle (SSDLC) and has expertise in security and
privacy frameworks such as NIST, MITRE, and GDPR.
His experience ranges from automated and manual testing to strategic consulting, including technical and executive
presentations. Additionally, he is fluent in Portuguese, English, and Spanish.
Author: Thiago Bispo
Check the schedule bsidesjp.com.br/#agenda
@BSidesJP
Daniel "d3v" Chactoura retweeted
PDF generators power a critical feature in almost every web application: invoices, reports, certificates, contracts. Behind that "Export PDF" button lies an often-forgotten attack surface. Combined, libraries like ReportLab, WeasyPrint, Dompdf, mPDF, Browsershot, ReLaXed, and TCPDF have over 90 million monthly installs.
This talk presents 9 real, security-critical vulnerabilities discovered in these widely-used libraries, from Remote Code Execution via deserialization and template injection, to Local File Read, SSRF, and Denial of Service. More importantly, you'll learn a practical 6-step methodology to identify and exploit vulnerabilities in any PDF export feature.
Key Takeaways:
- How to identify which PDF engine an application uses (and why it matters)
- Real exploitation techniques: HTML injection, SSRF bypasses, file oracles, and more
- A systematic approach to testing PDF generators in your next security engagement
- Defense strategies: what developers need to do right (and what they keep getting wrong)
Pedro Cruz is an Offensive Security Analyst at Hakai Security, specializing in penetration testing, vulnerability research, and bug hunting. He holds a degree in Information Security from FATEC Ourinhos and maintains a solid technical background backed by industry-recognized certifications, including OSCP+, CRTO, DCPT, and CHWI.
He works on executing Pentests (web, infrastructure, cloud, and mobile), generating technical and executive reports focused on real business risk. As an independent researcher, he uses AI to amplify the discovery of 0-days and misconfigurations in large-scale software, currently holding 4 published CVEs.
Author: Pedro Cruz @gankd_
BSides Zadar bsideszadar.org/ - Zadar , Croatia - September 11, 2026
@BSidesZadar
Daniel "d3v" Chactoura retweeted
During DefCon 34, the Hakai Security team presented and publicly released four new tools developed to enhance offensive security research, evasion, obfuscation, and mobile pentesting.
Check out the summary of what was just added to our GitHub:
Mirage: Agnostic network traffic reshaping in Rust to bypass statistical analysis and ML models. github.com/hakaioffsec/mirag…
MonProcessEX-Killer: Abuse of the MonProcessEX.sys kernel driver to terminate protected processes. github.com/hakaioffsec/MonPr…
Lyra: LLVM IR obfuscator for Rust that runs natively in cargo/rustc without modifying your code. github.com/hakaioffsec/lyra
BEERUS: Modular framework for complete lifecycle coverage of Android pentests. github.com/hakaioffsec/beeru…
Access all repositories via the link in the bio and share them with the community!
github.com/hakaioffsec
Daniel "d3v" Chactoura retweeted
Hakai Security dominated Las Vegas!
Our team had a massive presence on stage at BSides Las Vegas and DEF CON 34. We delivered 7 high-level technical talks, bringing cutting-edge research on exploit development, reverse engineering, detection evasion, mobile hacking, and APT tradecraft to the global cybersecurity community.
Seeing our team in the spotlight at some of the world's premier security conferences reinforces what we believe in: continuous research, innovation, and technical excellence in offensive security.
Congratulations to all the speakers for representing Hakai Security and Brazil at such a high level!
Daniel "d3v" Chactoura retweeted
How much memory should be required to change the tag of an NPM package? Just a few bytes, right?
During research conducted by Yokai, Hakai Offensive Security's research team, analysis of Gitea's Package Registry revealed that a seemingly harmless endpoint could be abused to cause a Denial of Service condition against the entire Gitea instance.
Gitea loaded the entire HTTP request body into memory before validating it, allowing an authenticated user to indirectly control how much memory the process allocated.
A single request with an excessively large body could terminate the Gitea process, while concurrent requests could keep the instance in a continuous cycle of crashes and restarts.
No administrative privileges, victim interaction, or access to other users' packages were required.
The vulnerability affected Gitea 1.26.2 and development versions prior to 1.27.0, and was fixed in Gitea 1.27.0.
CVE-2026-42931
Full technical analysis on the blog.
Check it out surl.li/thetxf
Author @_tricta





















