From SecurityWeek · 40 stories
Anthropic CEO Proposes Three-Step Plan to Slow AI Development Pace
Dario Amodei, CEO of Anthropic, has proposed a three-step plan to "pace the frontier" of AI development, aiming to slow its rapid advancement. The plan includes granting third-party evaluators access to AI models, establishing common industry safety standards with government involvement, and coordinating global regulatory efforts between democratic and authoritarian governments. This initiative seeks to address risks associated with AI and allow time for safeguards and regulation.
Pentagon Personnel Agency Data Breach Exposes 3 Million Records
The US Defense Manpower Data Center (DMDC) experienced a data breach exposing personal information for approximately 3 million individuals. Unauthorized users accessed a file-sharing server for nine months, compromising Social Security numbers, names, and other sensitive data.
Supply Chain Attacks Target Open Source, Impacting Over 2,500 Organizations
Software supply chain attacks targeting open source repositories and CI/CD systems have increased significantly, with a recent incident impacting over 2,500 organizations and 430,000 CI/CD pipelines. This attack, attributed to TeamPCP, initially compromised Aqua Security's Trivy vulnerability scanner and subsequently affected projects like LiteLLM, leading to the exposure of terabytes of credentials from major companies.
Azure Cosmos DB Vulnerability "CosmosEscape" Allowed Access to All Databases
Wiz Research discovered "CosmosEscape," a critical vulnerability in Azure Cosmos DB's Gremlin API that could have allowed attackers to compromise all databases within the service, including Microsoft's internal databases. The flaw enabled attackers to acquire a "Cosmos Master Key" for full read and write access. Microsoft has fully remediated the issue, eliminating the platform-wide key and adding new guardrails.
Jscrambler npm Package Supply Chain Attack Deploys Infostealer
The npm package Jscrambler version 8.14.0 was compromised, executing an infostealer on installation and affecting multiple subsequent versions. Released on July 11, 2026, the package was downloaded nearly 1,500 times before removal. The incident, attributed to credential compromise, highlights security risks in open-source dependencies.
N-able Takes N-central Offline After Researcher Discloses 33 Zero-Day Vulnerabilities
N-able has taken its N-central Hosted Environments offline and advised on-premises customers to shut down servers following the public disclosure of 33 zero-day vulnerabilities, collectively known as Able2Shell. These vulnerabilities, disclosed by security researcher Sean Bales, include credential disclosure, arbitrary file writes, and remote code execution. N-able is developing a hotfix, expected around November 10, and has not observed active exploitation.
Widespread Outages Affect Banks, Air Travel, NSA, and Apple; Not Cyberattacks
Major outages impacted the NSA, Apple, Microsoft Outlook, JPMorgan Chase, and the FAA over four consecutive business days. Director of National Intelligence Jay Clayton stated these disruptions were not due to cyberattacks or coordinated nation-state activity. Experts suggest the outages are a result of ongoing system updates and vulnerability patching, straining IT and security teams.
Nikkei Discloses Breaches of Employee Microsoft and Google Email Accounts
Japanese publishing company Nikkei reported two separate security incidents involving employee email accounts. Attackers accessed a Google Workspace account in July, potentially exposing personal information of 1,646 individuals, and a Microsoft 365 account in September, which was used to send 9,000 phishing emails to staff and interviewees.
Anthropic Releases Claude Fable 5.2 and Mythos 5.2 with Speed, Cost, and Capability Upgrades
Anthropic has released Claude Fable 5.2 and Claude Mythos 5.2, updated AI models designed for complex technical challenges. These models offer up to 30% faster output generation, reduced pricing to $8 per million input tokens, and improvements in coding, reasoning, and vision. Claude Fable 5.2 is generally available, while Claude Mythos 5.2 is accessible via verification programs for cybersecurity and life sciences.
Three AI Security Startups Raise Over $228 Million in Funding Rounds
Onyx Security, Obsidian Security, and Mindgard collectively raised over $228 million in recent funding rounds. Onyx Security secured $113 million for its AI agent control platform, Obsidian Security raised $85 million at a $1.1 billion valuation for agentic AI security, and Mindgard received $30 million for its AI security and red-teaming platform. These investments highlight increasing focus on securing AI systems and agents in enterprise environments.
New Mexico Court Orders Meta to Pay $25 Billion Penalty in Cambridge Analytica-Related Case
A New Mexico court has ordered Meta to pay a $25 billion penalty in a data privacy case stemming from the Cambridge Analytica scandal. This ruling follows a September jury finding that Meta misled state citizens about user data safeguards and the prevention of harmful content on its platforms, identifying 43 million violations of the state's unfair practices act. The penalty, initially calculated at $219.5 billion based on a maximum of $5,000 per violation, was reduced by a New Mexico judge, and Meta plans to appeal the decision.
Trump Dismisses AI Regulation Calls as 'Sick Conspiracy' Amid Industry Warnings
Donald Trump rejected calls for increased AI controls, labeling them a "sick conspiracy" and stating that a strong president is the only necessary "guardrail." This stance followed appeals from AI leaders like Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, Google DeepMind chief Demis Hassabis, and SpaceX boss Elon Musk for slower AI development, which contributed to a market downturn for AI-related tech stocks.
Google's Gemini AI autonomously hacked three companies in security test
Google's Gemini AI model autonomously breached three companies during a cybersecurity test, marking the first known instance of such an action. The AI found public information and guessed credentials to gain access, raising concerns about AI development and its potential for misuse.
Federal Agencies Broaden Alert on Iran-Linked OT Attacks Targeting More PLC Manufacturers
Federal agencies expanded an alert regarding Iran-affiliated hackers targeting internet-facing operational technology (OT). The updated warning now includes programmable logic controllers (PLCs) from Schneider Electric, Siemens, and potentially other manufacturers, beyond the previously identified Rockwell Automation and Allen-Bradley. This expansion highlights ongoing threats to critical infrastructure, emphasizing the need for secure PLC deployment and restricted internet access to prevent operational disruption and financial loss.
White House Authorizes Private Firms for Offensive Cyber Operations Against Foreign Cybercrime
The White House issued a presidential memorandum allowing vetted private U.S. companies to conduct offensive and intelligence-gathering cyber operations against foreign cybercrime organizations under federal control. This program, managed by the National Coordination Center, aims to counter transnational cyber threats and combat cybercrime, fraud, and predatory schemes by integrating private sector expertise into national security efforts.
Anthropic Releases Claude Opus 5, Offering Near Fable 5 Performance at Half the Cost
Anthropic has released Claude Opus 5, a new AI model that approaches the intelligence of its flagship Fable 5 model but at half the price. Opus 5 is now the default model for Claude Max subscribers and the strongest available for Claude Pro users, offering improved performance in coding and knowledge work while maintaining the token cost of its predecessor, Opus 4.8.
Three Charged in $129 Million AI Impersonation Scheme Targeting Global Companies
Three individuals have been indicted on charges of wire fraud and aggravated identity theft for using AI to impersonate CEOs and attorneys, stealing approximately $129 million from companies worldwide. The defendants mimicked voices and images of executives to trick accounting departments into wiring large sums, which were then converted to cryptocurrency.
Multiple Healthcare Data Breaches Impact Over 30 Million Individuals
Several healthcare organizations, including DentaQuest, Unlimited Technology Systems, MCBS, CareCloud, and Brown Health Medical Group-MA, have reported data breaches impacting over 30 million individuals. These incidents, occurring between May 2025 and March 2026, exposed sensitive personal, medical, and financial information, highlighting ongoing vulnerabilities in healthcare data security.
Unpatched Magento and Adobe Commerce Zero-Day Actively Exploited to Backdoor Online Stores
A new unpatched zero-day vulnerability, named StyleSmuggler, in Magento Open Source and Adobe Commerce is being actively exploited to install backdoors on online store servers. The flaw allows attackers to execute malicious code without authentication, affecting all current versions including 2.4.9. Adobe has not yet released a patch or advisory, leaving stores vulnerable to compromise.
Critical Rails Active Storage Flaw Allows Arbitrary File Read, Potential RCE
Ruby on Rails has patched a critical vulnerability, CVE-2026-66066, in its Active Storage framework that allows unauthenticated attackers to read arbitrary files from application servers. This flaw, with a CVSS score of 9.5, affects applications using libvips for image processing and accepting untrusted image uploads, potentially exposing sensitive data and leading to remote code execution.
Google Disrupts 1.3 Million-Device Android Botnet, Bans Manufacturer Videostrong
Google led a takedown of the XCPE botnet, comprising over 1.3 million Android devices, primarily TV set-top boxes. Google attributes the botnet to Chinese device maker Videostrong Technology Co., which allegedly intentionally embedded a backdoor in its products and sold access to criminals for ad fraud and proxy services. Google has terminated its relationship with Videostrong and banned the company from its Android ecosystem.
DHS Bans DeepSeek AI Use Across US Federal Civilian Agencies
The US Department of Homeland Security (DHS) has prohibited the use of DeepSeek AI products and models across all federal civilian agencies and their IT systems. This ban, effective since late July, stems from concerns that DeepSeek models could facilitate the transfer of sensitive US government data to the Chinese government and may contain malware or security vulnerabilities. The order also disallows DeepSeek from participating in federal government contracts.
OpenAI CEO Sam Altman to Attend Trump-Xi State Dinner in Washington
OpenAI CEO Sam Altman will attend President Trump's state dinner for Chinese President Xi Jinping's visit. His attendance occurs amidst ongoing discussions in Washington and Silicon Valley regarding AI regulation and the risks associated with advanced AI models.
OpenAI Introduces "Dots" Persistent AI Agents Powered by GPT-6 Astra
OpenAI launched "Dots" at its DevDay event, which are persistent AI agents built on GPT-6 Astra. These agents operate on their own cloud computers, use their own browsers, and can interact with over 4,000 applications through OpenAI's plugin ecosystem. Dots are designed to perform continuous tasks and automate workflows, available to ChatGPT Pro and Business Premium users.
Apple tightens macOS Full Disk Access controls due to AI agent risks
Apple is introducing new controls for macOS Full Disk Access settings, citing increased risks from AI agents. This change follows reports of AI apps potentially accessing private user data without explicit consent, aiming to ensure users understand the implications of granting such broad permissions.
Teens sentenced to 5.5 years for £29M Transport for London cyber attack
Owen Flowers and Thalha Jubair were sentenced to 5.5 years for a 2024 cyberattack on TfL that caused £29 million in damages. The attack severely disrupted services and breached data of millions. Authorities cite this case as a major enforcement action against young cybercriminals.
NSA's Classified Internal Network Experiences Outage During Routine Maintenance
The National Security Agency's internal classified communications network, NSANet, experienced temporary degradation during scheduled routine maintenance. The outage, which affected some internal systems and led to staggered personnel arrivals, was caused by applying a software patch for major PHP vulnerabilities. Mission-critical capabilities remained operational, and there is no evidence of a cyberattack.
Data Breach Exposes Information of Over 40,000 DHS Employees
A hacker group named "Department of Peace" published personally identifiable information for over 40,000 U.S. Department of Homeland Security (DHS) employees, including staff from U.S. Customs and Border Protection (CBP) and Immigration and Customs Enforcement (ICE). The data, which includes names, job titles, and work contact details, was reportedly stolen from an Okta single sign-on system used by DHS. Distributed Denial of Secrets (DDoSecrets) published the archive, named "Icepick," stating the goal was to replace "anonymity with accountability."
Dutch Police Arrest Man in Connection with ShinyHunters Hacking Group Investigation
Dutch police confirmed the arrest of a 24-year-old Amsterdam man earlier this month as part of an investigation into the ShinyHunters hacking group. The suspect, identified as Pepijn van der Stap, was previously arrested and sentenced for hacking and blackmailing multiple companies. This arrest is significant as it links a known individual to a prominent hacking group responsible for numerous data breaches.
Bitget crypto exchange reports $351.6 million stolen from hot and warm wallets
Cryptocurrency exchange Bitget announced that hackers stole $351.6 million from its hot and warm wallets. The company has suspended withdrawals and is investigating the incident, stating its User Protection Fund will cover all losses.
Google fined over €400m by Irish DPC for manipulating location data consent
Google received a fine exceeding €400 million from Ireland's Data Protection Commission (DPC) for manipulating users into agreeing to location data tracking. The DPC's six-year inquiry found Google lacked a valid legal basis for processing location data, which could reveal sensitive personal information.
Judge Rules Pentagon's Blacklisting of Anthropic as Supply Chain Risk Illegal
A San Francisco federal judge ruled that the Pentagon's designation of AI company Anthropic as a supply chain risk was illegal, citing a violation of the First Amendment. U.S. District Judge Rita Lin found the Department of Defense's actions were not based on an "articulable basis" and constituted "unlawful retaliation." This decision removes a barrier for Anthropic, potentially allowing defense contractors to use its technology.
WhatsApp Enhances Account Security with Stronger 2FA and Multiple Passkeys
WhatsApp has updated its account security features, allowing users to set alphanumeric passwords with special characters for two-step verification, replacing the previous six-digit PINs. The platform also introduced support for adding multiple passkeys to a single account, useful for users across iOS and Android devices, and added more context for calls from unknown numbers on Android.
TikTok settles DOJ children's privacy lawsuit for $400 million
TikTok and its parent company, ByteDance, have agreed to pay $400 million to settle a lawsuit filed by the U.S. Department of Justice (DOJ) in 2024. The lawsuit alleged violations of the Children's Online Privacy Protection Act (COPPA) by collecting personal information from underage users without parental consent. The settlement includes new age-related controls and parental oversight measures.
Cyberattack on CEVA Logistics Exposes European Steam Hardware Customer Data
A cyberattack on CEVA Logistics, Valve's European shipping partner, between July 29 and August 1, 2026, compromised personal data of Steam hardware customers in Europe. The breach exposed names, addresses, phone numbers, email addresses, and product details, leading Valve to warn customers about potential phishing attempts. This incident highlights supply chain vulnerabilities and impacts multiple retailers relying on CEVA Logistics.
Dependabot introduces default three-day cooldown for version updates
Dependabot now includes a default three-day cooldown before opening version update pull requests. This change aims to reduce the risk of merging compromised versions immediately after their release, enhancing supply chain security for developers.
Cybersecurity Expert Sentenced for Role in BlackCat Ransomware Scams
Angelo Martino, a former ransomware negotiator, has been sentenced to 70 months for aiding the BlackCat ransomware gang. Collaborating with accomplices, he shared confidential negotiation details, causing victims to lose over $75 million. This highlights vulnerabilities within cybersecurity industries.
Critical RCE Flaw in Forminator WordPress Plugin Affects 600,000+ Sites
A critical security vulnerability (CVE-2026-15748) in the Forminator Forms WordPress plugin allows unauthenticated attackers to upload malicious PHP files, leading to remote code execution and site compromise. The flaw, rated 9.8 on CVSS, affects over 600,000 active installations and requires specific form configurations for exploitation, but has been patched in version 1.56.2.
Ledger Wallets Sold by CryptoBillis Suspected of Hardware Tampering Leading to Crypto Thefts
Ledger is investigating reports of hardware-tampered wallets sold by reseller CryptoBillis, which have allegedly led to over $86 million in crypto thefts. The tampering involves a circuit board implant that intercepts seed passphrases and transmits them to attackers. This issue primarily affects users in Southeast Asia who purchased devices from CryptoBillis, with no indication of compromise to Ledger's direct sales or systems.
Chrome 150 Update Addresses 27 Vulnerabilities, Enhances Security
Google released Chrome 150, patching 27 security vulnerabilities, including two critical use-after-free flaws in Ozone and Views. This update is part of a broader effort to improve browser security, with most flaws discovered internally by Google. Regular updates are essential due to frequent exploitation of memory safety vulnerabilities in browsers like Chrome.