Últimas ameaças

3.131 casos · pág. 1/157

hackread.com · high

FBI warns of ARTEX AI attack framework used for privileged access in under 3 hours, incl. ConnectWise Cloud Backup supply-chain compromise

The FBI issued a Flash Advisory on November 13 warning that cybercriminals have used the open-source Go framework ARTEX since at least August 2026 against financial services, government, technology, defense, healthcare, semiconductor, energy, manufacturing, telecom and legal organizations worldwide. ARTEX connects to external LLMs (DeepSe…

blogs.cisco.com · low

Cisco publishes identity-centric security framework for agentic AI enterprises, citing survey of non-human identity incidents

This article is a Cisco blog post, not an incident report. It presents the Identity-Centric Security Framework for the Agentic Enterprise, a vendor-neutral reference architecture by Nancy Cam-Winget and Flemming Andreasen. The framework links agent execution, identity, authorization, policy, enforcement, connectivity, security intelligenc…

techcrunch.com · low

Jensen Huang says he would return as Nvidia CEO if the board allows, after resigning over affairs and China export-control scrutiny

Jensen Huang, founder of Nvidia, issued his first public statement since resigning. He said he would be honored to return as CEO if the board decides it serves shareholders, employees, the company and America. He acknowledged "serious mistakes" in his personal life and committed to cooperating with the board's independent investigation. A…

darkwebinformer.com · medium

Threat actor 'Chimera' claims leak of SCI Horizon database with 450,372 email records and 166,198 addresses

A threat actor using the handle Chimera claims to have leaked a database tied to SCI Horizon, described in the listing as a French commercial real estate company. The post advertises a 123 MB JSONL file of 450,372 lines, including 450,372 email records and 166,198 addresses. The visible sample shows user IDs, emails, activation status, ci…

theregister.com · medium

Fireworks AI breach exposes driver's license data of 149,346 Uber drivers via stolen supplier credentials

Uber confirmed that 149,346 people had personal data exposed after attackers breached its AI supplier Fireworks AI. Uber did not name the supplier in its filings, but told TechCrunch that it was Fireworks AI. The stolen data includes driver's license images and details extracted from them: names, home addresses, dates of birth and license…

securityonline.info · high

Java4Shell (CVE-2026-114938): Reported Unauthenticated RCE in OpenJDK 17 Amid Oracle's November 2026 Java SE Patch Update

Oracle's November 17, 2026 Critical Security Patch Update includes 11 Java SE fixes, all for flaws that may be exploitable over a network without authentication. One of them, CVE-2026-114938, has been nicknamed Java4Shell. Early reports describe it as unauthenticated remote code execution caused by memory corruption in OpenJDK 17 (17.0.6 …

helpnetsecurity.com · low

Linx Security launches identity controls for AI agents: shadow agent detection, runtime control, Autopilots, Prompt-to-Workflow

Linx Security announced four capabilities that extend identity governance to AI agent actions: Shadow AI Agent Detection, Agent Runtime Control, Autopilots and Prompt-to-Workflow. This is product news, not a threat or incident, and it names no vulnerability, attack or threat actor. The company says its customers already govern more than 2…

olhardigital.com.br · medium

Apple patches iOS/macOS audiomxd flaw (CVE-2026-65339) allowing microphone recording without user permission

A flaw tracked as CVE-2026-65339 in the audiomxd audio service of Apple iOS and iPadOS let an app record audio after the user denied microphone access. The app spoofed the identity of audiomxd through an audit token. This made the request look like an internal operation and bypassed the TCC (Transparency, Consent, and Control) check. In s…

Anthropic Cuts Coordinated Disclosure Deadline to 30 Days for High/Critical Bugs Found by Claude, Effective January 2027

Anthropic's Frontier Red Team announced on November 16 that it is shortening its coordinated vulnerability disclosure timelines for flaws its Claude models find and report to vendors and maintainers. The new rules apply to reports made on or after January 1, 2027. Reports already in progress keep their existing deadlines. Anthropic said t…