A vulnerability in Apple’s iCloud Private Relay tool allows malicious websites to expose users' real IP addresses, undermining its privacy protection. This issue affects all browsers on iOS due to flaws in Apple’s WebKit engine and also impacts the OnionBrowser app.

Apple's ‘Private Relay’ Is Exposing Users’ Real IP Addresses

404media.co

OpenAI banned a network of ChatGPT accounts linked to scams originating from Cambodia's Preah Sihanouk province. The network utilized ChatGPT for creating fake personas, generating scam messages, and managing operations across various schemes, including investment and romance scams.

OpenAI reveals how criminals used ChatGPT to run scams

helpnetsecurity.com

Forescout’s Vedere Labs identified 15 vulnerabilities in TP-Link's Omada routers, allowing attackers to hijack devices and intercept traffic. Exploiting predictable serial numbers, attackers can gain admin credentials and access internal networks.

15 TP-Link Omada vulnerabilities let attackers hijack routers and intercept camera traffic

helpnetsecurity.com

A cluster of 77 malicious extensions on Open VSX, uploaded between July 26 and August 1, 2026, has been removed as of August 3, 2026. These "evil twin" extensions impersonated legitimate tools, exfiltrating sensitive developer data to "mangorbit.com.

Open VSX Removes 77 Malicious Evil Twin Extensions Exfiltrating Developer Data

thehackernews.com

Cybersecurity researchers have identified a campaign, codenamed SMOKE#SCREEN, that uses fake Adobe and Zoom updates to deploy Remote Monitoring and Management (RMM) software, specifically ConnectWise ScreenConnect, for persistent remote access.

Fake Adobe and Zoom Updates Install ScreenConnect for Persistent Remote Access

thehackernews.com

The article discusses how adversaries are increasingly weaponizing AI for malicious purposes, as analyzed by Talos. It identifies three main categories of AI usage: creating sophisticated malicious software, scaling criminal operations, and conducting vulnerability research.

“Keep going, bro. You’ve got this!” A data-driven look at how adversaries are weaponizing AI

blog.talosintelligence.com

A supply chain attack has compromised multiple npm packages in the Keyv/Cacheable ecosystem due to a GitHub maintainer account breach.

keyv and cacheable npm Package Hijacked in Supply Chain Attack

wiz.io

The INC Ransomware operation is exploiting vulnerabilities in SonicWall SMA 1000 series VPN appliances, specifically CVE-2026-15409 and CVE-2026-15410, since August 2026. The group has claimed 885 victims, including organizations from Australia, the U.S., and Switzerland.

INC Ransomware Emerges as Dominant Actor Exploiting SonicWall SMA 1000 Flaws

thehackernews.com

Pillar Security researchers discovered a vulnerability in Google's Agent Development Kit for Python, allowing one AI agent to control another with higher privileges through prompt injection.

Google dev kit spurs first-ever agent-on-agent violence

theregister.com

Adobe has released security updates for Campaign Classic (ACC) addressing a critical flaw (CVE-2026-48449, CVSS 10.0) allowing arbitrary code execution without user interaction. Another high-severity flaw (CVE-2026-48448, CVSS 8.6) related to SQL injection was also fixed.

Adobe Campaign Classic CVSS 10.0 Flaw Could Run Code Without User Interaction

thehackernews.com

Apple is contesting a U.K. government demand for access to encrypted user data, filing a complaint with the Investigatory Powers Tribunal.

Apple challenges UK government’s latest demand for iCloud backdoor: report

techcrunch.com

This article discusses vulnerabilities in passwordless authentication, particularly focusing on Google’s synced passkey system.

The Police National Legal Database (PNLD) confirmed a data breach on July 26, exposing contact details of police, government, and customers on the dark web. Compromised data includes names and email addresses, potentially increasing phishing risks.

PNLD Breach Exposes U.K. Police and Government Contact Details on Dark Web

thehackernews.com

UK Government Investments (UKGI) experienced a data breach that exposed high-level management information and personal details of over 50 officials for nearly 40 hours. The breach was attributed to a staff member's failure to follow security protocols.

UK’s state investments agency hit by data breach

theguardian.com

Thermo Fisher Scientific has patched a critical vulnerability (CVE-2026-17583) in its Applied Biosystems human identification software, allowing undetectable alterations to DNA data files. The flaw, rated High with a CVSS score of 8.

Thermo Fisher Patches Flaw That Could Make DNA File Tampering Nearly Undetectable

thehackernews.com

Ruby on Rails has patched CVE-2026-66066, a critical vulnerability (CVSS score 9.5) in Active Storage that allows unauthenticated attackers to read arbitrary files and potentially achieve remote code execution. This affects applications using Active Storage with the libvips image processor.

Ruby on Rails Patches Critical Active Storage Vulnerability Affecting Image Processing

securityaffairs.com

An attacker exploited a firmware flaw in Coldcard hardware wallets, draining 1,082.65 BTC (approximately $70.2 million) from 1,196 addresses in 41 minutes on July 30. The issue stemmed from a March 2021 error that used a weak pseudorandom number generator for seed generation.

Coldcard Hardware Wallet Flaw Linked to $70 Million Bitcoin Theft in 41 Minutes

thehackernews.com

N-able disclosed a critical vulnerability (CVE-2026-18577) in its N-central RMM platform, allowing unauthenticated attackers to gain full administrative access. This affects all supported versions and is actively exploited. A hotfix (2026.3.1.7) was released on August 2, urging immediate upgrades.

Critical N-Able N-Central Vulnerability Allows Hackers to Gain god-mode Access to the RMM Console

cybersecuritynews.com

Security experts warn against using generic TV streaming devices, particularly H96 models, which are involved in ad fraud and residential proxy schemes. Research by Bitsight reveals these devices spoof mobile phone identities to click ads on AI-generated websites, generating significant revenue.

Read This Before You Buy That TV Streaming Stick

krebsonsecurity.com

Since early May 2026, Microsoft Threat Intelligence has tracked Storm-2945, part of Midnight Blizzard, executing targeted traffic manipulation attacks on hospitality networks globally.

CaptiveCrunch: Midnight Blizzard targets travelers worldwide for malware delivery and credential theft

microsoft.com

Unit 42 identified a Chinese-speaking threat actor using AI for autonomous cyberattacks, targeting infrastructure through seven vulnerabilities. The actor, known as knaithe and KnYuan, employed the Hermes Agent framework and DeepSeek for autonomous operations via Telegram.

Chinese-Speaking Threat Actor Harnesses AI Models for Autonomous Cyberattacks

unit42.paloaltonetworks.com

Google's Chrome browser is adapting to a surge in security vulnerabilities, with the Chrome security team reporting 1,072 fixes in two major releases in June, surpassing the total from the previous 23 releases.

Chrome Needs Twice-a-Week Patching Thanks to AI Bug Hunting

wired.com

Anthropic's Claude AI inadvertently accessed the live systems of three companies during cybersecurity testing due to a misconfiguration that left its environments online.

Anthropic’s Claude AI hacked three real companies during testing

theage.com.au

On June 23, 2026, Analog Devices, Inc. detected unauthorized access to its systems, activating incident response protocols and engaging external cybersecurity experts. The investigation revealed that certain files were exfiltrated, but the data has not been publicly released or used fraudulently.

The FTC has filed a lawsuit against telehealth provider Hims & Hers for allegedly sharing sensitive patient information with third-party advertising platforms, including Snap and Meta, despite claims of safeguarding privacy.

FTC sues Hims & Hers for allegedly sharing patient information with third-party platforms

therecord.media

Sweet Security announced its new Agentic AI Blocking capabilities on July 29, 2026, enhancing AI security by blocking rogue agent behavior in real time. This proactive runtime enforcement prevents unauthorized access to sensitive data and stops prompt injections before they can misdirect agents.

Sweet Security Brings Autonomous Protection to the AI Enterprise with New Blocking Capabilities

hackread.com

Wiz Research identified a critical vulnerability, CosmosEscape, in Azure Cosmos DB's Gremlin API, allowing attackers to access all databases, including Microsoft's internal ones. The exploit could retrieve the Cosmos Master Key, enabling full read/write access and enumeration of databases.

CosmosEscape: Taking Over Every Database in Azure Cosmos DB

wiz.io

A vulnerability in Microsoft Copilot allows attackers to embed malicious instructions in Word documents, which can alter document outputs and propagate the worm through normal workflows.

Word worm crawls into Copilot, spreads chaos

theregister.com

The average cost of a data breach has risen to $4.99 million, a 12% increase, according to IBM's 2026 Cost of a Data Breach Report, based on incidents from March 2025 to February 2026. Healthcare remains the most affected sector at $6.6 million.

The Average Cost of a Data Breach Rises to $5 Million

infosecurity-magazine.com

The U.S. Cybersecurity and Infrastructure Security Agency (CISA) added CVE-2026-20316, a zero-day vulnerability in Cisco Secure Firewall Management Center (FMC) Software, to its KEV catalog.

Cisco FMC Zero-Day Actively Exploited, Static Credentials Could Expose Sensitive Data

thehackernews.com