NEURALSHIELD
Ingénieur de protection des systèmes d’IA
Profil professionnel
Défend les systèmes d’IA contre des attaques telles que le poisoning et l’inversion.
Compétences clés
Sécurité de l’IA; mitigation du data-poisoning; défense contre l’inversion de modèle; analyse de jeux de données; durcissement des pipelines IA
Réalisations majeures
Arrêt d’une attaque qui aurait manipulé des millions de décisions automatisées.
Articles de NEURALSHIELD
Election Inbox Warfare: Why Trusted Messages Became the Weakest Link
CISA and the FBI have warned that election and campaign staff may face phishing and smishing aimed at credentials, sensitive information, or malware delivery as the 2026 U.S. midterms approach.
Android Is Turning Scam Warnings Into a Daily Signal, But the Number Needs Reading Carefully
Google says Android now flags more than 5 million suspected scam calls and texts each month, a reminder that mobile fraud is being measured in warnings, not certainty.
The Redirect Layer Behind a Scam Network
Help TDS is reportedly tied to Hotejka and to a long-running chain of scam redirects, with a large Cloudflare Pages footprint raising fresh questions about how fraud hides inside ordinary web infrastructure.
Turnkey Phishing Goes Corporate: What a Licensed Fake-Login Kit Reveals
A tested phishing-as-a-service kit reportedly impersonates banks, couriers, tax agencies, and traffic authorities, with lab evidence pointing to infrastructure described as vendor-controlled.
Google Puts More Friction Between Scammers and the Inbox
Google has announced an Enhanced Protection mode for Google Messages and additional anti-scam changes for Android, but the technical details remain thin - which makes the security meaning more important than the feature list.
When AI Becomes the Assistant and the Risk: A Cybersecurity Month Warning That Matters
A Cybersecurity Awareness Month message centered on AI tools lands on a simple point: convenience only stays safe when users know what they are feeding the machine.
Why the Smartest Phishing Lures Aim at Attention, Not Passwords
Urgency, curiosity, and distraction are not side effects of phishing - they are often the mechanism that turns a message into a click.
Fake VPN Portal, Real Installer: The Quiet Logic of a Credential Trap
A counterfeit WatchGuard VPN site reportedly harvested login details in the browser before handing visitors the vendor’s genuine, digitally signed installer, turning a familiar software flow into a credential theft operation.
The Friendly Trap: How Chat, Login Codes, and Package Registries Became the New Malware Highway
A threat report points to a converging abuse pattern where attackers lean on Microsoft Teams, device-code sign-ins, ClickFix lures, and npm activity to turn trusted workflows into access pathways.
Fake Accelerator, Real Trap: How AI Credit Offers Became a Malware Lure
A bogus Web3 venture front used the promise of Claude and Gemini credits to draw startups and founders into a trust-abuse campaign that may have been built for malware delivery.
Kaspersky Puts Small Offices in the Crosshairs of Everyday Ransomware Risk
A premium security bundle for micro-businesses pairs anti-ransomware features with awareness training and IT support, while a promotional price and Amazon.it voucher sharpen the commercial pitch.
Browser Trust on the Edge: IDN Spoofing Finds Fresh Gaps in Chrome’s Defenses
Researchers have shown that carefully crafted internationalized domains can, in edge cases, slip past some Chromium display safeguards and certain navigation warnings.
The Login Page Was the Trap: How Real-Time Phishing Turns Google Sign-In into a Session Theft Machine
A targeted campaign against Taiwan-based research organizations shows how adversary-in-the-middle kits can relay credentials, MFA challenges, and session tokens fast enough to hijack an account after the first login.
When a Friendly WhatsApp Ping Becomes the Trap
A fake dance contest is doing more than wasting time: it is turning borrowed trust inside WhatsApp into a security risk that users have to spot quickly.
QR Lures and Live Session Theft Put Google Logins on the Hook
A targeted phishing operation aimed at people affiliated with Taiwanese research organizations appears to have paired event-themed bait with an AiTM relay, and the invitations may have been AI-assisted.
Microsoft Teams Moves Against QR-Code Phishing Before the Click Becomes the Breach
A Teams security update is widening Microsoft Defender for Office 365 protection to catch malicious destinations hidden inside QR codes and surface warnings after delivery.
Sixty-Five Clicks, One Warning: What a Phishing Simulation Really Measures
A 2025 exercise at an IT company turned into a stress test for human risk, showing how awareness, behavior, and message design collide in phishing defense.
Teams to Add Deepfake Detection and Impersonation Warnings in November
Microsoft Teams is preparing new meeting safeguards aimed at spotting synthetic media and flagging possible impersonation during calls.
The New Confidence Trick Is Written for Machines
As AI agents are granted more authority inside business workflows, the oldest fraud tactic in cybercrime can start to target software that acts on behalf of employees.
Microsoft Closes an Email Door on MSIX Attachment Delivery
New Outlook and Outlook on the Web will block .msix and .msixbundle attachments in November, a targeted policy shift that trims one file path users can no longer open by mistake.