Bonjour à toutes et tous, Cette semaine, l’actualité cyber se concentre sur l’automatisation de l’extorsion, les fuites revendiquées et la pression exercée sur les organisations. Titan promet d’analyser les données volées et de calculer les rançons, tandis qu…

Deux jeux d’été mêlent fact-checking, détection IA et chiffrement pour entraîner l’esprit critique en famille.

Barbhack, DEF CON, Cyberbrew et Hackfest rythment la rentrée cyber, de Toulon à Québec, en passant par Lille.

Titan promet d’automatiser l’analyse des fuites, le calcul des rançons et la pression réglementaire.

B9 fait l’objet d’une fuite présumée de 36 000 profils, illustrant les risques cyber des banques 100 % en ligne.

Fuite Pokemon Gym : 19 600 comptes de joueurs et joueuses, adresses IP et messages privés exposés.

La police saisit les serveurs du site pour adultes Motherless dans une enquête européenne sur des contenus vidéos criminels.

Alphabet revendique des chiffres qui donnent le vertige : 119,8 milliards de dollars de revenus et un bénéfice net historique de 112,1 milliards. Mais les investisseurs se montrent de plus en plus frileux face à une facture IA qui ne cesse de grimper.

Lors d'un test en interne qui a mal tourné, deux modèles d'IA sur lesquels travaillaient OpenAI ont réussi à s'affranchir de leur confinement pour aller compromettre, en toute autonomie, des serveurs d'Hugging Face.

Bonne nouvelle pour les utilisateurs de la version gratuite de Gemini. Google vient de lancer les modèles Gemini 3.6 Flash et Gemini 3.5 Flash-Lite. La firme indique également qu’elle est déjà en train de tester Gemini 3.5 Pro, une IA très attendue qui devrai…

Anubis menace Coca-Cola après une cyberattaque revendiquée contre Fairlife et fixe un ultimatum au 27 juillet.

Coinbase Cartel menace Caterpillar, expose une stratégie d’extorsion cyber publique et alerte sur des pirates qui usurpent le groupe pirate !

Un escroc usurpe Damien Bancal et ZATAZ pour cibler les services communication par ingénierie sociale.

Un pirate affirme vendre des données présumés volés au Rassemblement national qui révélerait des données internes, sans preuve publique permettant d’authentifier la fuite à ce stade.

Une attaque autonome contre Hugging Face révèle les risques cyber des pipelines de données et des agents IA.

Stop the spread (of online recruiting and propaganda)

A threat actor used the open-source Hermes AI agent in unattended "YOLO" mode to automate post-exploitation activity during an alleged breach of Thailand's Ministry of Finance. [...]

Europol has flagged 4,340 URLs for removal during a multi-week operation targeting online content linked to "The Com," a loosely organized network of nihilistic violent extremist groups. [...]

Policy targets online scammers, sextortionists, and potentially their immediate families

Hunt.io uncovered a cyber-espionage attack on Thailand’s Finance Ministry using Hermes AI agent and Hades malware for reconnaissance and persistence. Researchers at Hunt.io have uncovered an intrusion targeting Thailand’s Ministry of Finance that offers a rar…

Industry professionals debate whether it represents a lab containment failure or an unprecedented agentic capability milestone. The post Industry Reactions to OpenAI Models Hacking Hugging Face: Feedback Friday appeared first on SecurityWeek.

An Illinois man was sentenced on Tuesday to 76 months in prison and three years of supervised release for hacking the Snapchat accounts of over 750 women to steal nude photos. [...]

The Computer Emergency Response Team of Ukraine (CERT-UA) has warned of a new campaign that involves the use of a malicious program that's dressed up as a Notepad++ plugin to compromise Windows systems. The activity has been attributed by the agency to a thr…

Origin Energy has confirmed that an unauthorized party accessed and subsequently leaked customer data online, exposing sensitive personally identifiable information (PII), among others. [...]

U.S. Cybersecurity and Infrastructure Security Agency (CISA) adds SharePoint and Check Point flaws to its Known Exploited Vulnerabilities catalog. The U.S. Cybersecurity and Infrastructure Security Agency (CISA) added DD-WRT, Langflow, and WordPress flaws to …

Phishing for dummies

Biggest overhaul since launch ditches sessions, guts little-used features, and leaves homebrew implementations facing a slog

Experts say the era of AI bug hunting is here, so defenders will simply have to adapt to busier workloads

Learn how cryptocurrency works, from blockchains and wallets to private keys, custody, and secure transactions, with practical security tips. for confident use.

CISA widens alert beyond Rockwell controllers as intruders target internet-facing devices across critical infrastructure

OpenAI's unintended attack on Hugging Face startled the world because its AI agent was acting on its own. But that's exactly what agentic AI is designed to do. We just didn't expect it to do it so well.

Stadler refuses $12.3 demand after thieves swipe technical data through supplier platform

Google DeepMind unveiled Gemini 3.5 Flash Cyber, an AI model for vulnerability discovery and patching, available only to governments and trusted partners. Google DeepMind announced Gemini 3.5 Flash Cyber on Tuesday, a security-focused AI model built on top of…

Google's new account recovery option prompts users to upload a selfie to prove their identity. Here's what to know.

The Chaos ransomware gang is using a new backdoor dubbed msaRAT that hides command-and-control (C2) communication by routing it through the Chrome or Edge browsers. [...]

Israeli cybersecurity firm Check Point Software has addressed an actively exploited zero-day flaw in the company's SmartConsole graphical user interface (GUI) admin panel. [...]

Closed models with guardrails can still cause harm, but may also not be able to fix problems they caused

Closed models with guardrails can still cause harm, but may also not be able to fix problems they caused

Beginning July 27, 2026, GitHub will cut public bug bounty payouts by at least half at every severity level. Critical findings will drop from $20,000-$30,000+ to a fixed $10,000, while its permanent invite-only VIP tier will pay $30,000 or more. Reports file…

OpenAI test model escaped on to internet & broke into a real company’s servers - posted in General Chat: https://www.cnn.com/2026/07/22/tech/openai-hugging-face-ai-cybersecurity   OpenAI says some of its experimental AI models left a test environment with no …

OpenAI models escaped from a controlled cyber test, exploited zero-day flaws and breached Hugging Face while searching its production database for test answers.

Enterprise AI can accelerate ransomware attacks when AI assistants and agents inherit excessive permissions or compromised identities. Acronis explains how identity controls, governance, and least-privilege access help reduce AI-enabled ransomware risk while …

Some never saw their files again either, infosec biz Proofpoint finds

The Cybersecurity and Infrastructure Security Agency (CISA) on Tuesday ordered U.S. government agencies to prioritize patching an actively exploited vulnerability in the Langflow visual framework for building AI agents. [...]

U.S. Cybersecurity and Infrastructure Security Agency (CISA) adds DD-WRT, Langflow, and WordPress flaws to its Known Exploited Vulnerabilities catalog. The U.S. Cybersecurity and Infrastructure Security Agency (CISA) added DD-WRT, Langflow, and WordPress flaw…

OpenAI confirmed its AI models exploited zero-days during internal testing, reaching Hugging Face servers in an unintended real-world cyberattack. OpenAI admitted on July 21 that its own AI models, including GPT-5.6 Sol and an unnamed pre-release system, were…

‘Digital transformation’ is a better bet, apparently, so private equity gets a turn

OpenAI says its AI models, including GPT‑5.6 Sol and a pre-release model, hacked into the Hugging Face artificial intelligence repository while being tested in a sandboxed testing environment. [...]

OpenAI on Tuesday said a combination of its artificial intelligence (AI) models, including GPT-5.6 Sol and an "even more capable pre-release model," was behind the security incident that targeted Hugging Face's production infrastructure last week. The AI com…

Sandboxed experiment found itself a zero day, escaped onto the open internet and stole data

Critical SharePoint RCE vulnerability CVE-2026-50522 is under active exploitation after the release of a PoC exploit code. A critical Microsoft SharePoint vulnerability, tracked as CVE-2026-50522 (CVSS score of 9.8), is being actively exploited following the …

New executive order calls for end-to-end visibility into defense supply chains, including software dependencies, foreign ownership and cyber-related supplier risks. The post Trump Orders Defense Contractors to Map Software, Suppliers Across Critical Supply Ch…

Hackers are exploiting the "wp2shell" critical vulnerability suite (CVE-2026-63030 and CVE-2026-60137) affecting WordPress Core to deploy persistent webshells and install malicious plugins on affected servers. [...]

Alleged developer arrested in Indonesia after more than 200 servers slain

Qilin ransomware exploits the PAN-OS GlobalProtect flaw CVE-2026-0257 to gain unauthorized VPN access to unpatched networks. Arctic Wolf researchers warn that the Qilin ransomware gang is exploiting the critical PAN-OS GlobalProtect vulnerability CVE-2026-025…

Google's DeepMind on Tuesday announced the release of Gemini 3.5 Flash Cyber, a specialized artificial intelligence (AI) model built atop 3.5 Flash that's designed to discover, validate, and patch vulnerabilities quickly and efficiently. According to the tec…

Attackers have begun to exploit two critical vulnerabilities in WordPress that, when combined together, enable unauthenticated remote code execution (RCE) and complete compromise of vulnerable websites. The two security flaws, tracked as CVE-2026-63030 and C…

Threat actors are now exploiting a recently disclosed critical security flaw impacting ServiceNow AI Platform, according to Defused Cyber. In a post shared on X, the threat intelligence firm said it's observing in-the-wild exploitation of CVE-2026-6875 (CVSS…

Attackers are exploiting critical ServiceNow flaw CVE-2026-6875, allowing unauthenticated remote code execution on self-hosted instances. Searchlight Cyber researchers disclosed a critical pre-authentication remote code execution vulnerability, tracked as CVE…