PATCHVIPER
Cavalier de patch pour systèmes industriels
Profil professionnel
Applique des correctifs en direct dans des environnements industriels où la production ne peut pas s’arrêter.
Compétences clés
Patchs sans interruption; sécurité OT; protection ICS; mitigation des vulnérabilités; sécurité de la production continue
Réalisations majeures
Application de correctifs en direct sur des machines industrielles, évitant un arrêt de production de 1,4 M€.
Articles de PATCHVIPER
When Reverse Engineering Gets an AI Co-Pilot, the Real Risk Moves to the Plumbing
A new open-source project is wiring AI coding agents into established binary-analysis tools, turning the Model Context Protocol into a fresh control point for defenders and attackers alike.
One File-Read Bug, One Identity Bridge, One Dangerous Shortcut to Jira Control
A critical Atlassian flaw shows how a narrow file-access issue can become far more serious when self-hosted software is tied to Crowd and sensitive credentials sit on disk.
Autonomous Red Teaming Gets Fresh Capital as Defense Turns to Speed
Hadrian’s $40 million raise points to a simple cyber reality: defenders are increasingly buying automation that can validate exposure faster than manual testing cycles allow.
Crash First, Escape Later? VMware PoC Puts Guest-to-Host Boundaries Under Pressure
A public proof-of-concept for a critical VMware flaw has so far produced a host crash, but the bigger concern is the security line it tests between a virtual machine and the machine running it.
VMware’s Guest Boundary Takes a Hit as a Public PoC Lands
A critical VMware bug tied to VMXNET3 can let a privileged guest turn a virtual machine into a path toward host code execution, and the public PoC raises the pressure to patch fast.
Pwn2Own Ireland Turns Zero-Days Into a Measurable Risk Map
The contest ended with $1,262,000 in rewards and 98 zero-day flaws demonstrated, offering a rare look at where modern systems are most likely to crack first.
Fuzzing’s Real Job: Finding the Bug Before the Bug Finds You
A 2026 tooling roundup is less about vendor names than a basic security truth: the fastest way to reduce software risk is to hit code with bad inputs before users do.
Zammad’s Double Fault: When a Helpdesk Bug Becomes Host Takeover
A reported chain of two previously unknown Zammad flaws shows how a support platform can turn from business utility into a route to full server control.
A Console Web Page, a Kernel Bug, and the Thin Line Between Research and Jailbreak
A newly published proof-of-concept for PS5 firmware 7.00 through 13.60 shows how a browser-engine foothold can be chained into kernel-level read/write access.
The Hidden Safari Tag That Can Rebuild a Browsing Story
A small artifact in Safari’s History database can help investigators infer browsing themes, but only when it is read alongside URLs, timestamps, cache data, downloads, and network telemetry.
Safari History Can Reveal More Than Pages: The Hidden Context in macOS Trails
A little-known Safari artifact may help investigators see broader browsing themes on macOS, even when the page list is incomplete or unclear.
AI Is Shrinking the Exploit Window, Not Rewriting Security
The pace of vulnerability hunting is changing, but the defensive answer is still the same old one: layered controls, disciplined patching, and virtual patching only as a temporary bridge.
AI Security Tools Are Moving Into AWS - And the Real Battle Is Over Trust
A new autonomous vulnerability discovery service keeps source code inside customer AWS environments, but the harder problem is proving that AI findings are accurate, contained, and actually fixed.
API Security Is Fracturing Into Use Cases, and That Changes How Defenders Buy
A 2026 roundup of eight API security tools points to a practical shift: discovery, pre-release testing, and runtime defense are being treated as separate jobs, not one bundle.
Bug Bounty Platforms Are Becoming a Trust Decision, Not a Brand Contest
Choosing a vulnerability program is increasingly about fit: maturity, jurisdiction, and access control matter more than a familiar logo.
PS5 Exploit Chain Draws a Straight Line From Browser Code to Kernel Memory Control
A newly named chain, Relapse, illustrates how a web-facing weakness can become a much deeper system problem when it is paired with a kernel memory bug.
When the Request Lies: Why Host Headers and Redirects Still Break Web Trust
A new technique-focused post puts two old web bugs back under the microscope: host header injection and open redirects, both of which matter because applications often trust request data too quickly.
When Branch Predictors Turn into Memory Leaks
A newly disclosed Spectre-v2 style technique shows how reused JIT code can confuse CPU prediction state and, in a proof of concept, expose Linux secrets including a root password hash.
The Analytics Gate That Should Have Stayed Shut
A reported authentication flaw in Microsoft’s Titan analytics service is a reminder that in data platforms, the login layer can be as sensitive as the data itself.
When Scanning Is Not Enough: The Push Toward Validated Exploits in AI Pentesting
AI-assisted penetration testing is being used to separate a vulnerability that exists from one that can actually be exercised in a controlled test.