Security · Top stories
Savannah Police Fire Six Employees for Misusing License Plate Reader System
The Savannah Police Department terminated four officers and two civilian employees for misusing the Flock Safety License Plate Reader system. An internal investigation found employees conducted unauthorized searches on personal acquaintances and family members, and one officer granted unauthorized access to an external agency.
Deveillance Launches Spectre I Device to Counter AI-Enabled Wearable Recording
Tech startup Deveillance introduced Spectre I, a device designed to prevent unauthorized recording by AI-enabled wearables. This development responds to the increasing prevalence of AI accessories that can silently record conversations and activities, raising privacy concerns.
Hardware Backdoor Discovered in Some x86 CPUs, Allowing Kernel-Level Access
Researchers have uncovered a hardware backdoor, dubbed "rosenbridge," in certain desktop, laptop, and embedded x86 processors, specifically VIA C3 CPUs. This backdoor allows userland code to bypass processor protections and freely read and write kernel data, and has been found enabled by default on some systems.
REpsych Toolset Generates Images from Program Control Flow Graphs
Christopher Domas released REpsych, a proof-of-concept toolset that translates source images into functioning programs where the program's control flow graph (CFG) generates the original image. This project demonstrates a novel method of manipulating program visualization, primarily for reverse engineering tools.
Stade Français restores systems after cyberattack, investigates data leak
French rugby club Stade Français Paris confirmed a cyberattack disrupted its information systems, but has since restored operations from backups. The club is investigating a data leak after a sample of allegedly stolen data was published online by the Qilin ransomware group. This incident highlights the increasing targeting of sports organizations by financially motivated hackers.
Nepal's National Cyber Security Centre Joins Have I Been Pwned Government Service
Nepal has become the 47th government to join Have I Been Pwned's free government service, granting its National Cyber Security Centre access to monitor Nepalese government domains for data breaches. This integration allows Nepal to identify exposed government email addresses and respond to compromised accounts, strengthening its cybersecurity posture.
Trump raised cyber scam compound issue with Xi, State Department official says
A State Department official informed senators that former U.S. President Donald Trump discussed Southeast Asian cyber scam compounds with Chinese President Xi Jinping. This issue is significant due to the billions of dollars siphoned from Americans annually by these operations, often run by Chinese organized crime groups.
Georgia Investigates Alleged Foreign Disinformation Campaign Targeting Russian Tourists
Georgia's State Security Service has initiated a criminal investigation into an alleged foreign disinformation campaign. This campaign reportedly used fake social media accounts to spread fabricated stories portraying Georgia as unsafe for Russian tourists, aiming to damage the country's national and economic interests.
Ship Safe: Open-source security scanner for code, AI agents, and supply chains released
Ship Safe, an open-source security scanner, has been released to identify vulnerabilities in application code, AI agents, and supply chain components. This tool operates locally within repositories, offering features like AI-backed red-teaming and automated remediation, which helps developers find and fix security issues before deployment.
DuckDuckGo iOS App Adds Copy Clean Link Feature to Remove Tracking Identifiers
DuckDuckGo's iOS app introduced a new "Copy Clean Link" feature that automatically removes tracking identifiers from URLs when copied. This update allows users to share links without inadvertently transmitting personal tracking data to recipients.
SIEMatic, a fair-sourced observability and security platform, released as early-access alpha
SIEMatic, a new fair-sourced observability platform for security event analysis, has been released in an early-access alpha state. This platform provides tools for collecting, indexing, searching, and analyzing security events, offering an alternative for security monitoring and incident response.
Metro Bank customer loses £14,000 in AI chatbot-linked fraud despite bank notification
A Metro Bank customer lost over £14,000 to fraudsters who purchased credits for the AI chatbot Claude, even after he notified the bank that an initial suspicious transaction was illegitimate. The bank failed to fully freeze his account, allowing subsequent fraudulent payments to proceed. This incident highlights ongoing security challenges related to AI service payments and bank fraud detection systems.
River Bank Confirms Stolen Data Deleted After Ransomware Attack
River Financial Corporation confirmed that data stolen during a June 16 ransomware attack was deleted by the threat actor. The company engaged with the hackers to ensure data deletion, likely involving a ransom payment, but has not confirmed if personal information was exfiltrated or the attack's material impact.
Nightcrawler: A Local AI Pentesting Agent Running on a Smartphone
Nightcrawler is an autonomous penetration testing agent that operates entirely on a smartphone, using a local AI model to discover hosts, map services, find vulnerabilities, and generate reports without cloud connectivity. This tool automates security assessments directly from a mobile device, offering a stealthier approach compared to traditional vulnerability scanners.
PISIGuard Browser Extension Protects Sensitive Information in AI Chats
PISIGuard, a new browser extension, automatically detects and masks personally identifiable and sensitive information (PISI) before it is sent to AI chat platforms like ChatGPT, Claude, and DeepSeek. This tool addresses privacy concerns by ensuring sensitive data, such as names, emails, and API keys, is replaced with placeholders during AI interaction and restored in the AI's response, with all processing occurring locally in the browser.
Netflix Sued After Unencrypted Movie Drive Stolen From Office Desk
Netflix is being sued by Op-Fortitude and financier Simon Afram after an unencrypted hard drive containing the unreleased movie "Fortitude" was stolen from a Netflix office desk. The lawsuit alleges Netflix failed to implement basic security measures and delayed informing the production company about the theft, potentially leading to a leak and financial losses.
UK Police Program Diverts Teen Hackers Towards Cybersecurity Careers
The UK's Cyber Choices program, run by police, is working with teenage hackers to redirect their skills towards legitimate cybersecurity paths. This initiative aims to prevent young individuals from engaging in illegal cyber activities by offering guidance and training opportunities.
Certisfy Introduces Cryptographically Signed URLs for Trust Verification
Certisfy launched a new feature allowing users to publish cryptographic signatures as claims for URLs, enabling verification of link providers. This functionality aims to establish trust and authenticity for shared links, particularly in an online environment with AI-generated content.
Cantina Exits Stealth with $8 Million Funding for Autonomous Vulnerability Management
Cybersecurity startup Cantina has emerged from stealth mode, announcing $8 million in new funding, bringing its total raised to $16.5 million. The company offers an autonomous security platform that uses AI agents to identify, prioritize, and remediate software vulnerabilities. This development introduces another player in the evolving field of AI-driven cybersecurity solutions.
Ariana Grande Sues Unidentified Hackers for Leaking Unreleased Music and Videos
Ariana Grande has filed a lawsuit in Los Angeles County Superior Court against unidentified hackers who allegedly stole and leaked her private content, including 45 unreleased songs in 2023 alone. The lawsuit aims to uncover the identities of these individuals and addresses the ongoing issue of content theft targeting artists and their collaborators.
Hulios: An eBPF-powered, transparent Tor gateway for Linux released
Hulios is a new transparent Tor VPN gateway for Linux that uses eBPF to redirect network traffic through Tor, securing outbound connections. It operates at the kernel socket level to prevent IP leaks and does not modify system configurations or rely on external firewalls. This tool provides a new option for Linux users seeking enhanced privacy and leak protection for their network traffic.
Arizona Resident Protests City's Use of Flock Cameras with Satirical Surveillance Plan
Jacob Petrosky presented a satirical plan to surveil government officials to the Casa Grande City Council, mirroring the city's use of Flock cameras on its citizens. This act was a protest against the city's $10 million contract with Flock for surveillance technology, which Petrosky has opposed multiple times.
Flashpaper Launches Self-Destructing Secret Sharing Tool with No Database
Flashpaper, a new open-source tool, offers self-destructing secret sharing without using a database, storing data in RAM only. It provides browser-based zero-knowledge encryption and an API for AI agents, though the API flow is not zero-knowledge. This tool provides a new option for ephemeral information exchange, particularly for users prioritizing data deletion and minimal server-side data retention.
OT Security Startup Frenos Raises $1.52 Million in Seed Round Extension
Frenos, an AI-native operational technology (OT) security startup, secured an additional $1.52 million in a seed funding round extension, bringing its total raised to $6.4 million. This funding will support the expansion of its customer success team and AI research and development, and coincides with the launch of its SAIRA Co-Work AI reasoning agent.
Romanian Air Force Shoots Down Third Drone in Three Days Over National Airspace
The Romanian Air Force shot down a third unauthorized drone on Sunday, July 26th, following two similar incidents on Friday and Saturday. These events indicate repeated airspace violations near the Ukrainian border, prompting defensive actions by the Romanian military.
Individuals disabling Flock Safety surveillance cameras due to privacy concerns
Individuals are actively disabling Flock Safety automated license plate reader (ALPR) cameras, citing concerns over mass surveillance. This movement highlights growing public resistance to widespread deployment of ALPR technology and its implications for privacy.
Abstract Security Raises $25 Million to Expand Composable Security Operations Platform
Abstract Security secured $25 million in a new funding round, bringing its total funding to nearly $50 million, to advance its streaming-first security operations platform. This investment will be used to enhance in-stream detection, expand workflow capabilities, and grow its go-to-market team, impacting the development of composable security solutions.
VaultSort's Guardian feature uses local AI to identify sensitive files on Mac
VaultSort introduced Guardian, a new feature that scans Mac computers for sensitive files such as identity documents, financial records, and credentials. This tool operates entirely on-device using AI, ensuring no user data is uploaded to external servers or cloud APIs, which addresses privacy concerns for sensitive information.
Webinar Highlights Behavioral AI for Modern Email Security Threats
BleepingComputer is hosting a webinar on July 8, 2026, focused on the limitations of traditional email security against evolving threats like phishing and business email compromise. Experts will discuss how behavioral AI can automate the detection and response process, addressing attacks that leverage trusted identities and legitimate workflows.
Apple's iPhone Lockdown Mode: Features, Limitations, and Target Users
Apple's Lockdown Mode, available since iOS 16, enhances security against spyware by disabling several iPhone features. This mode is intended for a small group of users, such as journalists and activists, who face targeted commercial spyware threats, rather than the average user.
Silent Patches Blind Defenders While Still Exposing Vulnerabilities to Attackers
Vendors who quietly fix vulnerabilities without public advisories or CVEs do not prevent attackers from discovering the flaws. This practice primarily hinders defenders, including penetration testers, vulnerability management engineers, and IT administrators, by denying them crucial information needed to assess and prioritize patches.
AI's Impact on Cybersecurity: Increased Attack Sophistication and Defender Challenges
Generative AI has significantly altered the cybersecurity landscape by enabling more sophisticated attacks, making traditional methods like phishing easier to execute, and introducing new threats like voice cloning and deepfakes. This development presents a challenge for cybersecurity defenders who must now contend with AI-powered adversaries.
Residential Proxies Posed as National Security Threat, Calls for Government and ISP Action
The use of residential proxies is identified as a national security threat due to their role in enabling social media manipulation, spam, and covert influence campaigns. The article advocates for government intervention and ISP action to detect and mitigate these proxy networks, which often operate through compromised user devices.
Integrating Autonomous AI and Analyst Copilots in Security Operations Centers
A Fortune 50 CISO discussed AI agents in security operations centers (SOC), revealing limitations in current designs that may overlook many alerts requiring human judgment. The insights of psychologist Daniel Kahneman emphasize the need for a balanced approach between automatic and deliberate human cognition in AI architecture for effective security.
Bug Bounty Programs Face Challenges from AI-Generated Submissions, Quality Improves for curl
Some bug bounty programs, like curl's, have faced issues with declining submission quality due to AI-generated reports, leading to program pauses or closures. However, curl's program has seen an improvement in submission quality, with confirmed vulnerabilities returning to previous levels, despite continued AI use in reports. This highlights the importance of verifying findings for aspiring bug hunters.
Safe Remediation Method Addresses Delays in Applying Urgent Security Fixes
A method called "Safe Remediation" is proposed to address delays in applying urgent security fixes by providing evidence of changes before implementation. This approach aims to build confidence among operations teams that security actions will not disrupt services. It involves four stages: Observe, Prove, Act, and Confirm, using existing data and testing to validate proposed changes.
Analysis of US Election Security Focuses on Voting Machine Risks and AI Threats
This analysis examines the cybersecurity aspects of US election integrity, specifically focusing on voting machine vulnerabilities and the potential for AI to influence elections. It details the process of in-person voting with ballot-marking devices and discusses how foreign influence campaigns aim to undermine public trust.
Security Questionnaires Remain Central to Vendor Due Diligence Despite Known Weaknesses
Security questionnaires are a critical but flawed component of vendor due diligence, often delaying deals and consuming significant GRC team resources. Companies rely on these self-reported assessments despite recognizing their limitations, as they serve as the primary gatekeeper for business transactions.
Attackers Operate in Gaps Between Security Alerts, Exploiting SOC Operational Limits
Attackers exploit the time between security alerts and blend into normal operations to evade detection, making it difficult for SOC teams to trace the full scope of an intrusion. Modern attacks use AI for speed but also employ stealth tactics like slow reconnaissance and using standard admin tools to remain undetected for extended periods.
Robot Vacuums Pose Privacy Risks Due to Data Collection and Security Vulnerabilities
Robot vacuums collect sensitive household data, including floor maps and potentially audio/video, which can be transmitted to manufacturers' servers. A recent incident involving a DJI Romo vacuum exposed a backend permission flaw, allowing one user to control thousands of devices globally, highlighting security weaknesses beyond typical user credentials.