2026-07-16
Population: stories the tracker first saw on this Coordinated Universal Time (UTC) calendar day; publisher date is used only if first-tracked time is unavailable and remains on each item.
- ransomware
Coca-Cola says Fairlife ransomware attack halts US dairy production
Coca-Cola disclosed that a ransomware attack on its Fairlife dairy subsidiary has disrupted US operations and temporarily halted production of Fairlife products. The attack impacts supply chains for a major dairy brand sold nationwide.
Why it matters: Retailers and distributors relying on Fairlife products face inventory disruptions; enterprises should review incident response protocols for critical supply chain partners and monitor ransom demands or data exfiltration claims.
- government policy
Senator calls on Rubio, Blanche to push back against Canadian surveillance legislation
Democratic Senator Ron Wyden called on the Trump administration to pressure Canada against adopting proposed surveillance legislation. Wyden contended that the measure would repurpose American technology infrastructure for surveillance objectives.
Why it matters: U.S.-based technology companies and infrastructure operators face risk if Canada implements surveillance laws that leverage their systems; policymakers should monitor diplomatic efforts on this issue.
- ai security
Agentic AI Is Untamable: Ask the Right Security Questions
The article notes that agentic artificial intelligence introduces new risks that extend beyond traditional attacker threats. It argues that organizations must reconsider their security approaches to address these challenges.
Why it matters: Organizations deploying agentic AI face uncertain behavior that may bypass existing controls, prompting a review of AI governance and threat models.
- breaches incidents
The Breach That Won’t End: An Update on Canvas, and how they created an EdTech’s Vendor Trust Problem
Instructure's forensic investigation into its Canvas breach continues to extend beyond initial timelines, with the company only beginning to deliver institution-specific findings in early July after months of delay. Educational institutions are receiving data packets through secured file-sharing links as Instructure completes its review of customer-specific impacts.
Why it matters: Canvas administrators and information security teams need to assess breach findings promptly upon receipt and identify what user data was exposed at their institution to determine notification and remediation obligations.
- ai security
1M+ Emails Use Hidden Text to Dupe AI Security Filters
Over one million emails are using hidden text techniques to evade artificial intelligence-based security filters. Text salting, a method that obscures content while remaining readable to humans, proves effective at bypassing AI-driven email defenses and allowing malicious messages to reach user inboxes.
Why it matters: Email administrators and security teams rely on AI filters to catch phishing, but this evasion technique puts all users at risk of delivery of deceptive messages that bypass those protections.
- threat intel
Now, even Russia's most elite hackers are using Clickfix to infect devices
Sandworm, an elite hacking unit within Russia's GRU military intelligence, has adopted the Clickfix technique to target sensitive organizations in Ukraine. Clickfix tricks users into copying and pasting malicious commands into a terminal by masking them as CAPTCHA verification steps, leading to malware installation or data theft. Ukrainian authorities discovered at least 10 compromised websites using this method, with confirmed infections from FreakyPoll, a custom Sandworm malware package.
Why it matters: Organizations in Ukraine and those handling sensitive data are at immediate risk from state-sponsored attacks using a low-friction social engineering technique; security teams should educate users to never paste unknown commands into terminals and monitor for Clickfix lures.
- vulnerabilities
Claude Chrome extension flaw lets malicious extensions trigger AI actions
A flaw in Anthropic's Claude for Chrome browser extension allows malicious extensions to trigger predefined artificial intelligence (AI) actions through simulated user clicks. An attacker exploiting this vulnerability could potentially abuse Claude's access to connected services including Gmail, Google Docs, Google Calendar, and Salesforce.
Why it matters: Chrome users relying on the Claude extension face credential and data exposure risk if a malicious extension gains foothold; practitioners should audit extension permissions and patch when an update becomes available.
- government policy
Program to rotate cyber personnel through federal agencies saw little use
The Federal Rotational Cyber Workforce program, launched in 2022 to shuffle skilled personnel among federal agencies, enrolled only eight participants across 634 applications over its three-year run. The Government Accountability Office attributed the low uptake to declining advertised positions, applicant qualification issues, and agencies preferring internal rotations, while the Office of Personnel Management cited budgetary constraints and has discontinued the program as of summer 2026.
Why it matters: Federal security leaders must recognize that workforce development initiatives struggle without sustained funding and clear participation incentives, limiting the federal government's ability to build cyber expertise across agencies during a critical shortage period.
- threat intel
New OkoBot framework deploys 20 payloads to steal data, crypto
OkoBot is a newly identified malicious framework delivering over 20 payloads in coordinated attacks targeting cryptocurrency wallets, credentials, and sensitive data theft. The framework appears designed for large-scale data exfiltration across multiple attack vectors.
Why it matters: Security teams protecting cryptocurrency users and enterprises handling sensitive credentials need to identify and block OkoBot payloads; the multi-payload architecture suggests established distribution infrastructure.
- ransomware
Ransomware and Cyber Extortion in Q2 2026
A ransomware threat report covering Q2 2026 activity found that group rankings shifted significantly, with "The Gentlemen" claiming the top spot while previously dominant groups like Qilin and DragonForce declined. Notable developments include Deadlock's resurgence with blockchain-based command and control and kernel-level endpoint detection and response (EDR) evasion techniques, alongside steady targeting patterns favoring professional, scientific, and technical services sectors across 90 groups in 99 countries. Despite quarterly fluctuations in victim counts, underlying attack techniques remained consistent, with the US absorbing approximately 49 percent of victim activity.
Why it matters: Security teams must prioritize detection of ransomware behaviors (remote-service abuse, identity compromise, lateral movement, defense evasion) rather than tracking group rankings, and should immediately assess exposure to Deadlock's blockchain-based command and control and kernel-level EDR termination tactics that may evade current defenses.
- threat intel
Russian trio indicted for allegedly running bulletproof hosting providers that spurred cybercrime
Three Russian nationals were indicted in 2024 for operating bulletproof hosting providers, Media Land and ML.Cloud, that supported cybercriminals conducting attacks on critical infrastructure across 21 U.S. states and multiple countries, resulting in losses exceeding $62 million. The defendants allegedly provided malware and ransomware distribution infrastructure, technical support for phishing and brute-force attacks, and hosted criminal marketplaces. The U.S. Treasury Department and allied governments imposed sanctions on the defendants and their companies in November 2025.
Why it matters: Critical infrastructure operators in any of the 21 affected states and allied nations should investigate whether their systems were targeted, and all practitioners should understand that bulletproof hosting providers remain a key enabler of ransomware and extortion campaigns that directly threaten operational continuity.
- ransomware
The Real Cost of Ransomware Is What Cripples the Business
A new analysis reveals that financial executives significantly underestimate ransomware expenses, typically by a factor of ten or more. The article examines the actual cost calculation methodology and explains why standard risk assessments fail to capture the true financial impact of ransomware incidents.
Why it matters: CFOs and security leaders need to recalibrate ransomware risk valuations in their risk registers today, as current estimates may understate exposure by an order of magnitude and inform inadequate resource allocation for prevention and response.
- breaches incidents
Italy fines WINDTRE €1.7 million over data breaches
Italy's data protection authority fined telecommunications operator WINDTRE 1.7 million euros for inadequate data security that led to two unauthorized breaches affecting over 365,000 customers. The regulator cited serious shortcomings in the company's security systems as the basis for the enforcement action.
Why it matters: Telecom operators and any organization handling customer personal data face regulatory exposure for insufficient security controls; practitioners should audit their breach notification processes and remediation timelines against local data protection requirements.
- regulatory
UK investigates TikTok for alleged age-verification lapses, exposing kids to online harms
Ofcom, the UK's communications regulator, is investigating TikTok for potentially inadequate age-verification measures that may have exposed children to harmful content. The regulator emphasized that age checks are fundamental to compliance with UK online safety requirements and indicated that many services lack sufficient verification systems.
Why it matters: UK-based businesses and platform operators subject to online safety regulations must ensure robust age verification; TikTok faces potential enforcement action if lapses are confirmed, signaling regulatory priorities for child protection compliance.
- vulnerabilitiesCVE-2026-32201CVE-2026-45659
CVE-2026-32201, CVE-2026-45659, CVE-2026-56164: Frequently Asked Questions About Active Exploitation of Microsoft SharePoint Server Vulnerabilities
Cybersecurity and Infrastructure Security Agency (CISA) confirmed active exploitation of four Microsoft SharePoint Server vulnerabilities, including CVE-2026-32201, CVE-2026-45659, CVE-2026-56164, and CVE-2026-58644, which attackers chain together to compromise on-premises deployments. The flaws enable spoofing, remote code execution (RCE), and privilege escalation, with attackers using them to steal IIS machine keys and establish persistence. Microsoft has released patches for all five vulnerabilities in the alert, with detection signatures available for three of them.
Why it matters: Organizations operating SharePoint Server on-premises (versions 2016, 2019, and Subscription Edition) must patch immediately to block active in-the-wild attacks exploiting these critical flaws for unauthorized access and malware deployment.
- government policy
Ukrainians rally against dismissal of tech-minded defense minister Fedorov
Ukraine's President Volodymyr Zelensky dismissed Defense Minister Mykhailo Fedorov, who had championed the integration of drone technology and digital innovation into the military. Fedorov's removal prompted public opposition from Ukrainians who supported his technology-focused defense initiatives.
Why it matters: Defense organizations globally watch Ukraine's military innovation approach; dismissal of a tech-forward defense leader may signal shifts in priority for cyber and digital warfare capabilities that other nations are actively developing.
- identity access
Least privilege for AI agents: Identity, access, and tool binding
Artificial intelligence (AI) agents that operate across multiple systems and invoke tools in sequences introduce identity and authorization challenges when deployed without managed identities and least-privilege role-based access controls (RBAC). Organizations are provisioning AI capabilities faster than their identity models can evolve to constrain them safely, leading to risks including unauthorized data access, unintended modifications, and potential privilege escalation from overly broad role assignments. The article recommends treating each agent as a first-class principal with lifecycle-managed identity, explicit role assignments, tightly scoped permissions, and preconfigured tool manifests.
Why it matters: Security teams and platform engineers need to implement least-privilege AI agent identity and access controls today to prevent scope creep in permissions, unauditable cross-system actions, and incidents that auditors and customers cannot fully explain.
- ransomware
ThreatsDay: Game Cheat Spyware, 24-Hour Ransomware, Chrome Sync Stalking + 12 More Stories
This week's security news covers multiple attack vectors including game cheat spyware, rapid ransomware deployment, and Chrome synchronization exploitation for surveillance. The stories highlight how attackers leverage seemingly legitimate tools, weak configurations, and straightforward attack paths to compromise systems and escalate damage.
Why it matters: Practitioners need to monitor for threats using familiar-looking repositories and installers, review sync settings across browsers and tools, and patch or mitigate known attack chains that remain effective despite public disclosure.
- research
Period tracker Stardust shares users’ health data with analytics firm, says Mozilla research
Mozilla research on period tracker applications revealed significant privacy disparities among tested apps, with one application sharing users' health data with an analytics firm while another demonstrated strong privacy protections. The findings highlight inconsistent data handling practices in the period tracking app ecosystem.
Why it matters: Users of period tracking apps face unexpected data sharing to third parties; practitioners should assess which apps in their organization or recommendations contain these privacy gaps and consider privacy-first alternatives.
- ot ics
Legacy Systems, Real-World Impacts: The Reality of OT Security
An article examining the challenges of balancing vulnerability disclosure in operational technology environments where legacy systems, safety concerns, and critical infrastructure risks complicate security practices. The piece addresses the tensions inherent in securing aging systems that cannot always be quickly patched or updated.
Why it matters: OT security practitioners and asset owners need to understand disclosure best practices when legacy infrastructure cannot be rapidly remediated, as improper handling of vulnerabilities in these environments can directly impact physical safety and continuity of critical services.
- industry
How a former DeepMind researcher raised at a $300M pre-seed valuation before launching a product
A former DeepMind researcher who contributed to foundational work on systems underlying ChatGPT has raised $300 million at pre-seed valuation before launching a product. The founder believes visual artificial intelligence (AI) represents a major frontier for the field's next phase of development.
Why it matters: Practitioners tracking emerging AI capabilities and startup momentum should monitor this founder's technical direction, as prior contributions to large language model research may signal novel approaches to vision-based AI systems.
- threat intel
Early Attack Warning Signals for AI and Agentic Security
Silent Push released a Model Context Protocol (MCP) Server that feeds Context Graph intelligence directly into Artificial Intelligence (AI)-assisted investigation tools. The tool lets practitioners query adversary infrastructure data in plain language inside compatible platforms such as Claude and Cursor. By surfacing Indicators of Future Attack (IOFA) weeks before weaponization, it helps security teams act on early warnings.
Why it matters: Security operations and threat intelligence teams gain early visibility into adversary infrastructure, allowing them to block attacks before they launch.
- ai security
Protecting Privacy in an AI Era
Daniel Solove argues in the Wall Street Journal that individual control of personal data is insufficient for privacy regulation in the artificial intelligence (AI) era. Instead, he proposes holding companies accountable through data minimization, fiduciary duties, liability for negligent design and harmful algorithms, and multi-stakeholder technology review. This approach parallels regulatory models used for food and drug companies.
Why it matters: Organizations handling personal data need to understand that regulatory pressure is shifting from individual consent models to corporate accountability and algorithmic liability, affecting how they design systems and manage risk.
- ai security
AI Agents Broke the Security Playbook. Here's What Replaces It.
Traditional security workflows were designed for slower-moving environments and do not adapt well to artificial intelligence (AI) agents operating at machine speed. Token Security argues that effective AI agent security requires building on live identity foundations and allowing security teams to customize workflows for their specific contexts. The shift reflects a fundamental mismatch between legacy security automation and the pace and complexity of AI-driven operations.
Why it matters: Security teams adopting AI agents need to rethink detection, response, and access control processes today, or risk gaps between their playbooks and how AI systems actually behave in production.
- vulnerabilities
Demystifying AI Exploits: A Blueprint for AI-Assisted Vulnerability Management
Mandiant reports that the mean time-to-exploit has fallen to negative seven days, prompting security teams to integrate large language model agents into vulnerability management workflows. The guidance emphasizes establishing operational guardrails, such as data security, workload isolation, and zero data retention agreements, to mitigate risks like prompt injection and privilege escalation. Frameworks like the National Institute of Standards and Technology Artificial Intelligence Risk Management Framework and Google’s Secure Artificial Intelligence Framework provide structural baselines for safe deployment.
Why it matters: Security teams using or planning to use AI agents for vulnerability management must implement guardrails to prevent data exposure, privilege escalation, and abuse of cloud provider policies.
- threat intel
Sandworm hackers have a CAPTCHA trick for Ukrainians
Sandworm, a Russian state-sponsored hacking group, is distributing a malicious CAPTCHA that tricks Ukrainian users into executing a PowerShell command on their Windows systems. The fake verification appears legitimate but actually delivers malware when users follow the embedded instructions.
Why it matters: Ukrainian organizations and individuals are at direct risk of compromise through this social engineering attack; security teams should warn users to verify CAPTCHA legitimacy and avoid copying unfamiliar commands from web forms.
- vulnerabilities
Adaptiva simplifies secure patch management for air-gapped networks
Adaptiva released AirGap for OneSite Patch, a new feature that applies autonomous patch management to air-gapped networks, allowing isolated systems to receive security updates while maintaining physical separation. The capability targets government agencies, critical infrastructure operators, and enterprises that require high-security isolation standards.
Why it matters: Organizations managing air-gapped networks can now patch critical systems without breaking isolation requirements, reducing exposure to vulnerabilities in disconnected environments.
- vulnerabilities
n8n Token Exchange Flaw Could Let Attackers Log In as Users From Another Issuer
n8n, a workflow automation platform, contained a JWT validation flaw in Enterprise instances configured with multiple external token issuers. The vulnerability allowed an attacker with a valid token from one issuer to gain unauthorized access to another user's account by exploiting weak claim matching that validated only the subject (sub) claim while ignoring the issuer (iss) claim.
Why it matters: Enterprise customers using n8n with multiple federated identity providers face account takeover risk; teams should audit token issuer configurations and apply fixes immediately.
- ai security
AI Data Centers Are Being Built Faster Than They Can Be Secured
Security teams note that the rapid rollout of Artificial Intelligence (AI) infrastructure is outpacing the deployment of traditional protective controls. Existing data‑center designs were not created to secure the novel workloads, interconnects, and hardware accelerators that characterize AI environments. As a result, organizations face increased risk of misconfigurations and exploitable weaknesses until security architectures evolve to match the speed of AI deployment.
Why it matters: Security and infrastructure teams building AI data centers face heightened risk of misconfigurations and must prioritize security controls that keep pace with rapid deployment.
- vulnerabilities
Sunsetting the Public AttackerKB Platform
Rapid7 will retire the standalone AttackerKB website on August 18, 2026, consolidating its vulnerability intelligence and exploit analysis into the Rapid7 blog, Vulnerability and Exploit Database, and customer-specific application programming interface (API) offerings. The public AttackerKB API and community contribution model will be discontinued to maintain data quality and reduce the risk of inaccurate or manipulated vulnerability information. Existing content will remain accessible through redirects and new technical write-ups will continue on the Rapid7 blog with a dedicated tag.
Why it matters: Practitioners relying on the public AttackerKB API or platform must migrate to Rapid7's centralized sources by August 18 or lose access to that intelligence feed; those using the website will be redirected to the Vulnerability and Exploit Database.
- vulnerabilities
The best defenders build AI agents together: Join Tenable for Swarm at Black Hat ’26
Tenable is hosting Swarm, a hands-on build event at Black Hat 2026 where security practitioners will collaborate to develop open-source artificial intelligence (AI) agents and tools for collective defense. Gartner forecasts that Fortune 500 enterprises will deploy over 150,000 AI agents by 2028, yet most security teams currently build agents in isolation without shared learnings or coordinated impact. The event aims to unite practitioners to create next-generation agents, skills, and model context protocol (MCP) servers that streamline security operations and threat response.
Why it matters: Security teams adopting agentic AI need visibility into shared, battle-tested open-source tools and collaborative practices to avoid duplicating effort and maximize the operational gains from AI-driven automation across their organizations.
- threat intel
‘ClickLock Stealer’ Bypasses macOS Security With Social Engineering, Process Killing
A macOS malware called ClickLock Stealer has targeted at least 100 users by combining social engineering tactics with process killing techniques to bypass security protections and steal passwords and cryptocurrency. The attack method exploits user interaction to circumvent macOS defenses.
Why it matters: macOS users and organizations supporting Mac endpoints need to recognize this threat uses social engineering to appear legitimate, making traditional security controls insufficient without user awareness training.
- threat intel
New ClickLock macOS Stealer Kills Apps Every 210ms Until Victims Type Their Password
ClickLock Stealer is a new macOS infostealer that arrives via a Terminal command and displays a fake system dialog requesting the user's login password. When users decline, the malware installs LaunchAgents that repeatedly terminate core system applications like Finder, Dock, and Spotlight every 210 milliseconds until the victim relents and provides their credentials.
Why it matters: macOS users and administrators need to block Terminal-based installation vectors and educate staff to refuse suspicious password prompts, as this stealer renders systems unusable to coerce credential theft.
- ransomware
Scattered Spider members behind TfL hack get five years in prison
Two prominent members of the Scattered Spider cybercrime group received sentences of five years and six months each for conducting a hack against Transport for London in 2024. The sentences represent judicial action against key figures in the criminal collective responsible for the intrusion.
Why it matters: Organizations targeted by Scattered Spider and security teams tracking the group should monitor this case outcome as it signals law enforcement capability to prosecute major cybercriminals, while the group's remaining members may adjust tactics or targets in response.
- threat intel
Russian cybercriminal used jailbroken Gemini CLI to rebuild botnet infrastructure in six minutes
Between March 19, 2026 and April 21, 2026, a Russian-speaking actor using the handle "bandcampro" employed a jailbroken Gemini CLI to set up a small command-and-control (C2) network. The actor used this infrastructure to control eight computers in a dental clinic and to access the clinic’s OpenDental database. TrendAI reported that the activity spanned over 200 sessions during the period.
Why it matters: Dental clinics running OpenDental face exposure of patient records if threat actors abuse jailbroken artificial intelligence (AI) agents to create command-and-control (C2) infrastructure; security teams should audit AI tool usage and watch for unexpected command-and-control traffic.
- vulnerabilitiesCVE-2025-11698CVE-2025-12011
Rockwell Automation CompactLogix, ControlLogix, Compact GuardLogix and GuardLogix
Rockwell Automation has disclosed three denial-of-service vulnerabilities (CVE-2025-12011, CVE-2025-12012, CVE-2025-11698) affecting multiple CompactLogix, ControlLogix, Compact GuardLogix, and GuardLogix controller models with a CVSS score of 8.6. These flaws could allow a remote attacker to load an invalid project file, causing affected devices to enter a major non-recoverable fault. Rockwell recommends updating to patched versions such as V35.016, V36.011, or later depending on the controller model.
Why it matters: Organizations operating critical manufacturing infrastructure worldwide using these Rockwell Automation controllers are exposed to production downtime if systems are exploited; patching to the recommended versions is required to restore service availability and prevent operational disruptions.
- vulnerabilitiesCVE-2026-54798CVE-2026-54799
Siemens SICAM 8
Siemens released a security advisory for multiple SICAM 8 products affected by four vulnerabilities, including active debug code exposure, firmware signature validation flaws, insecure default configurations, and unverified password changes. The vulnerabilities range from CVSS 6.5 to 7.2 and could enable denial of service, malicious firmware installation, and unauthorized system access. Siemens recommends updating to firmware version 26.20 or later.
Why it matters: Critical infrastructure operators in manufacturing and energy sectors worldwide using SICAM 8 components (A8000, EGS, S8000) must prioritize patching to prevent attackers from disrupting operations, compromising system integrity, or gaining unauthorized control of essential functions.
- vulnerabilitiesCVE-2026-11889
SALTO ProAccess Space
SALTO ProAccess Space versions before 6.13 contain CVE-2026-11889, a privilege escalation vulnerability that allows authenticated attackers to bypass authorization and access spaces outside their assigned partition when the tenancy feature is enabled. The vulnerability requires valid operator credentials and does not affect installations without partitioning. SALTO recommends upgrading to version 6.13 and implementing network isolation, least-privilege access controls, and consideration of separate instances for strong tenant separation.
Why it matters: Organizations operating SALTO ProAccess Space with partitioning enabled in critical infrastructure (commercial facilities, critical manufacturing) worldwide should upgrade to 6.13 immediately to prevent authorized insiders from accessing unauthorized areas.
- vulnerabilitiesCVE-2026-9292
Rockwell Automation FactoryTalk DataMosaix
Rockwell Automation FactoryTalk DataMosaix Private Cloud versions 8.02 and earlier contain a stored cross-site scripting (XSS) vulnerability that allows authenticated attackers with high privileges to inject malicious scripts into the Workflows configuration. The vulnerability could enable account takeover, credential theft, or malicious redirection when other users access affected pages. Rockwell Automation recommends upgrading to version 8.03 or later to remediate the issue.
Why it matters: Organizations running DataMosaix Private Cloud 8.02 or earlier in manufacturing and IT environments worldwide should upgrade immediately, as internal attackers with elevated access can compromise other user accounts and systems through script injection.
- vulnerabilitiesCVE-2026-15352
NASA Core Flight System (cFS) Health & Safety (HS) Application
A null pointer dereference vulnerability (CVE-2026-15352) in NASA's Core Flight System Health & Safety application can cause a denial-of-service condition when processing Housekeeping Telemetry requests. The flaw affects versions prior to v7.0.1, which NASA recommends as the remediation. CISA scored the vulnerability as High severity with a CVSS v3.1 base score of 7.5 and notes no known public exploitation at this time.
Why it matters: Organizations operating NASA cFS HS application in critical infrastructure or space systems need to update to v7.0.1 immediately to prevent denial-of-service attacks that could disrupt mission-critical telemetry and health monitoring functions.
- vulnerabilitiesCVE-2026-57896CVE-2026-60063
AutomationDirect Productivity Suite
AutomationDirect Productivity Suite versions up to 4.6.2.2 contain multiple vulnerabilities including out-of-bounds write, out-of-bounds read, and divide-by-zero flaws that require local or physical access to exploit. These issues can lead to memory corruption, information disclosure, application instability, or denial-of-service conditions. The vendor recommends updating to version 4.7.0.47 or later, with network isolation and access controls as interim mitigations.
Why it matters: Critical manufacturing facilities worldwide running affected versions face privilege escalation and system instability risks; practitioners should prioritize patching or implementing network isolation immediately if updates cannot be deployed.
- vulnerabilitiesCVE-2026-9653
Rockwell Automation 1756-EN2, 1756-EN3, and 1756-ENBT
Rockwell Automation disclosed a denial-of-service vulnerability (CVE-2026-9653) affecting communication modules 1756-EN2, 1756-EN3, and 1756-ENBT used in critical manufacturing environments worldwide. The flaw stems from improper validation of CIP Implicit Connection packets, allowing network-based attackers to send crafted packets that temporarily disrupt device connections. Patches are available for the EN2 and EN3 modules (version 12.002), though the ENBT module is discontinued with no fix available.
Why it matters: Manufacturing and critical infrastructure operators deploying these Rockwell Automation modules face service disruption risk if exposed to the public internet or untrusted networks; immediate patching of EN2 and EN3 devices and network segmentation of ENBT modules are required to prevent availability attacks.
- regulatory
Windows 11 24H2 Home and Pro reach end of support in 90 days
Microsoft will end support for Windows 11 24H2 Home and Pro editions in 90 days, ceasing all security updates for these versions. The affected systems will no longer receive patches or maintenance after the end-of-support date.
Why it matters: Organizations and individuals running Windows 11 24H2 Home or Pro must plan upgrades or migrations before support ends to avoid operating unpatched systems exposed to known vulnerabilities.
- threat intel
20+ Hijacked Government Websites Became an Attack Channel
Researchers discovered that more than 20 Brazilian government websites were compromised and used to distribute malware as part of an active campaign called PhantomEnigma. The investigation identified previously unknown backdoor functionality, infrastructure links, and multiple attack components associated with the campaign.
Why it matters: Security teams protecting government networks and endpoints need to identify and remediate any artifacts from these compromised sites, and defenders should monitor for indicators of compromise (IOCs) from this campaign across their infrastructure.
- breaches incidents
AU: Regulator’s preliminary findings did not indicate Qantas breached privacy obligations
Australia's Office of the Australian Information Commissioner completed a preliminary inquiry into the Qantas data breach affecting 5.67 million customer records and determined no likely violation of privacy obligations occurred. The investigation examined how the 2025 incident unfolded after an attacker gained unauthorized access to the airline's systems.
Why it matters: Qantas customers and Australian organizations should note that regulatory clearance does not necessarily indicate adequate security practices; practitioners should review their own incident response, disclosure, and data protection measures regardless of preliminary findings.
- cloud saas
ValorC3 extends SaaS protection with immutable cloud backups
ValorC3 Data Centers released a fully managed backup service for SaaS applications including Microsoft 365, Entra ID, and Salesforce, featuring immutable backup copies to protect against deletion, corruption, and ransomware. The service addresses a common misconception that cloud vendors automatically provide data backup protection. Recent governance research indicates 80% of organizations have encountered at least one cloud security incident.
Why it matters: SaaS users relying on Microsoft 365 and Salesforce should evaluate whether their native backup capabilities meet recovery needs from ransomware or accidental deletion, as shared responsibility models leave data protection gaps.
- breaches incidents
Thalha Jubair and Owen Flowers sentenced to prison
Two individuals, Thalha Jubair (age 20) and Owen Flowers (age 18), were sentenced to five years and six months in prison at Woolwich Crown Court for hacking Transport for London. The judge indicated the motivation behind their actions was primarily criminal in nature. This case represents enforcement action against unauthorized access to critical infrastructure systems.
Why it matters: Organizations operating public transportation and critical infrastructure should reinforce that attacks on these systems carry serious criminal penalties, and should review access controls and logging to detect similar intrusions early.
- ai security
New Agent Data Injection Attack Can Make AI Agents Misclick or Run Attacker Commands
Researchers have identified a data injection attack that manipulates artificial intelligence (AI) agents by corrupting the information they consume from external sources, allowing attackers to redirect actions without hijacking the underlying task. A planted fake review can cause a shopping agent to make unintended purchases, and a fake code comment can trick a coding assistant into executing arbitrary commands. The attack succeeds because agents trust the data they retrieve and act on corrupted information while continuing their original objective.
Why it matters: Organizations deploying AI agents for customer-facing tasks, code automation, and information retrieval face immediate risk of fraudulent transactions, system compromise, and data exfiltration; practitioners should evaluate how their agents validate external data sources before acting on user instructions.
- industry
Oak Emerges From Stealth Mode With $60 Million in Funding
Oak, an identity management startup, has completed a $60 million funding round and launched from stealth mode. The company has developed an artificial intelligence (AI)-powered Identity Operating System designed to manage and govern identities across an organization's entire environment.
Why it matters: Security teams and identity administrators should evaluate Oak's platform if they manage sprawling identity infrastructures across multiple systems and need centralized AI-driven governance.
- threat intel
Daxin Resurfaces in Taiwan Alongside Stupig Pre-Login SYSTEM Backdoor
A China-linked threat actor has redeployed Daxin, a kernel-mode rootkit inactive for over four years, in a Taiwan manufacturing company alongside a new backdoor named Stupig. Daxin was previously documented by Symantec in March 2022 as a tool used in targeted attacks.
Why it matters: Manufacturers in Taiwan and other sectors face renewed risk from sophisticated state-sponsored malware; security teams should assess whether Daxin or Stupig exist in their environments and review kernel-mode persistence mechanisms.
- vulnerabilities
CISA orders feds to patch actively exploited Oracle flaw by Saturday
The Cybersecurity and Infrastructure Security Agency (CISA) has issued a mandatory directive requiring federal agencies to patch a critical vulnerability in Oracle E-Business Suite that is currently being exploited in active attacks. The deadline for remediation is Saturday.
Why it matters: Federal agencies running Oracle E-Business Suite must immediately identify and patch affected systems to comply with CISA's order and defend against active exploitation.
- vulnerabilities
Splunk, Zoom Patch Critical Vulnerabilities
Splunk and Zoom released patches for critical vulnerabilities that could enable attackers to obtain credentials, access data, take over user accounts, and elevate privileges.
Why it matters: Organizations running Splunk or Zoom should prioritize applying these patches to prevent credential theft, account takeover, and privilege escalation attacks against their infrastructure and users.
- breaches incidents
Romania’s land registry hit by cyber attack, data allegedly for sale
Romania's National Agency for Cadastre and Land Registration experienced a cyber attack on July 14 that took down its e-Terra cadastre and land registry application. Initially reported as a technical incident, the disruption has been confirmed as an attack, and the agency states that data has not been compromised, though the investigation remains ongoing.
Why it matters: Organizations managing critical land registry infrastructure and practitioners in government IT security should assess their incident response protocols, as attackers targeting cadastral systems can compromise property records and enable fraud across real estate and financial sectors.
- threat intel
Russian hackers trojanize WebEx, Zoom apps to push Starland malware
A Russian financially motivated threat actor identified as UAT-11795 is distributing trojanized versions of WebEx and Zoom applications to deliver Starland RAT, a new backdoor designed to steal credentials and cryptocurrency.
Why it matters: Organizations and individuals using WebEx or Zoom are at risk of credential theft and financial loss if they download software from untrusted sources; practitioners should validate application authenticity and monitor for Starland RAT indicators.
- ai security
AI Can Find Bugs, But Human Knowledge Still Proves Them
Artificial intelligence can accelerate offensive security work by rapidly reading code, generating payloads, and automating repetitive testing tasks. However, human expertise remains essential to validate findings and confirm that AI-identified potential bugs are genuine vulnerabilities before they become actionable.
Why it matters: Security teams considering AI-assisted tools should expect to staff human validation and proof-of-concept work; AI augments speed but does not eliminate the need for skilled analysts to confirm results.
- ai security
The Hunter's Paradox: Is it time to embrace automated threat hunting?
The article discusses the tension between the need for automation in threat hunting due to growing data volumes and the lack of full trust in artificial intelligence that can be misled by attacker deception. It argues that while humans cannot keep pace with telemetry, AI systems often accept misleading inputs at face value, which skews their judgments. The author proposes that resolving the paradox starts with revising the definition of threat hunting to incorporate both human insight and automated checks.
Why it matters: Security analysts and threat hunters face delayed detection if they rely only on manual hunts or unchecked AI, so they should validate automated outputs before acting.
- threat intel
UAT-11795 deploys novel Starland RAT and bespoke WLDR C2 implant in financially motivated campaign
Cisco Talos identified UAT-11795, a Russian-speaking financially motivated threat actor conducting campaigns since at least June 2025 against victims in the U.S. and Europe. The group deploys Starland RAT, a Python-based remote access tool, alongside WLDR, a sophisticated PowerShell-based command-and-control implant featuring encrypted beaconing and task queuing capabilities. The actor distributes trojanized installers of legitimate software such as MobaXterm, WebEx, Zoom, and DBeaver to establish persistence and steal credentials and cryptocurrency assets.
Why it matters: Organizations and users in the U.S. and Europe face credential and cryptocurrency theft risk from trojanized downloads of common developer tools and collaboration platforms, requiring immediate review of installation sources and endpoint detection tuning for Starland RAT and WLDR beaconing patterns.
- ransomware
New Spirals ransomware encrypts victim network in under 24 hours
A previously unknown ransomware actor named Spirals has demonstrated the ability to compromise a corporate network and complete encryption within a single day. The speed of the attack, from initial access through data theft to full encryption, indicates a highly efficient operational capability.
Why it matters: Security teams need to accelerate detection and response procedures, as this threat actor can move from entry to encryption in less than 24 hours, leaving minimal time for manual intervention or escalation.
- vulnerabilities
Tenable One unifies code risks with enterprise exposure data
Tenable announced that its One Exposure Management Platform now incorporates static code vulnerability data alongside existing exposure information. The integration provides security teams with continuous visibility from code commit through runtime across the entire attack surface. By doing so, Tenable aims to help organizations address vulnerable code before it reaches production, especially as artificial intelligence (AI)‑generated code accelerates development cycles.
Why it matters: Security and DevSecOps teams gain a unified view of code and runtime risks, enabling them to prioritize and remediate vulnerabilities faster in environments where AI‑generated code speeds up release cycles.
- vulnerabilities
Unpatched Shark Vacuum Flaw Could Let Attackers Control Other Vacuums Region-Wide
A researcher disclosed an unpatched vulnerability in Shark RV2320EDUS robot vacuums that allows extracting a certificate from one device to gain root access to other vacuums operating on the same AWS region. An attacker exploiting this flaw could control cameras, movement, access home maps, and retrieve plaintext Wi-Fi passwords from affected devices.
Why it matters: Owners of Shark RV2320EDUS vacuums and other models using the same architecture face immediate risk of unauthorized device control and network credential theft; practitioners managing IoT device security should assess exposure and contact Shark for patched firmware.
- vulnerabilities
F5 Patches Multiple NGINX, BIG-IP Vulnerabilities
F5 released patches for multiple vulnerabilities affecting NGINX and BIG-IP products. The flaws could enable attackers to modify configurations, terminate processes, cross security boundaries, leak memory, and execute arbitrary code.
Why it matters: Organizations running F5 NGINX or BIG-IP should prioritize patching to prevent configuration tampering, denial of service, privilege escalation, information disclosure, and remote code execution in critical infrastructure.
- ai security
Lineation.ai focuses on runtime security for autonomous AI agents
Lineation.ai launched a runtime security platform designed to protect autonomous artificial intelligence (AI) agents during execution. The platform combines a zero trust control plane with an endpoint daemon to monitor agents that access sensitive data, call application programming interfaces (APIs), and execute workflows, addressing gaps in traditional perimeter and large language model (LLM) gateway defenses.
Why it matters: Enterprises deploying autonomous AI agents face new attack surface from agents with data access and execution privileges; practitioners need runtime visibility and controls to prevent compromised or manipulated agents from misusing permissions.
- government policy
China’s Top Cybersecurity Firms Hit by Mounting Military Procurement Bans
Chinese cybersecurity firms are experiencing procurement restrictions from the country's military, though specific reasons for the bans remain undetailed in available reporting. The action reflects tensions within China's defense-industrial ecosystem regarding vendor relationships or compliance frameworks.
Why it matters: Organizations relying on Chinese cybersecurity tools for supply chain integration should monitor potential shifts in vendor viability and support continuity, as military procurement restrictions may signal regulatory or geopolitical pressures affecting broader market participation.
- identity access
Microsoft makes Windows SSO prompts easier to manage
Microsoft introduced a new registry-based policy allowing IT administrators to automatically accept Windows single sign-on (SSO) permissions on Windows 11 versions 24H2 and 25H2 devices managed through Microsoft Entra ID. Users with personal Microsoft accounts and unmanaged devices will still receive SSO permission prompts. The change addresses European Economic Area (EEA) regulatory requirements for user choice in the sign-in experience.
Why it matters: Administrators managing Entra ID-joined Windows 11 devices can now reduce friction in SSO workflows through policy configuration, streamlining user authentication while maintaining controls for personal accounts and unmanaged endpoints.
- threat intel
Police take down investment fraud network that stole €100 million a month
Dutch and Belgian police, with Europol support, dismantled a transnational investment fraud network operating approximately 20 call centers with over 700 employees posing as financial advisers. The organization generated an estimated €100 million monthly by targeting victims across multiple countries, with the primary suspect, a 46-year-old with Israeli and Polish citizenship, taken into custody.
Why it matters: Financial services firms, compliance teams, and fraud detection units should monitor for call center-based investment scams targeting their customers and review authentication controls to prevent impersonation of legitimate advisers.
- industry
VS Code agent host runs Copilot, Claude, and Codex in a dedicated process
Visual Studio Code 1.129 introduces a dedicated agent host process that runs artificial intelligence (AI) coding agents such as Copilot, Claude, and Codex using the Agent Host Protocol. This architecture isolates each agent session in its own process, enabling developers to manage multiple AI coding tools without maintaining separate editor windows.
Why it matters: Development teams using multiple AI coding assistants should evaluate whether the dedicated process isolation improves stability, performance, or security posture for their workflows.
- ransomware
Srsly Risky Biz: Ransomware Uses AI To Amp Up Negotiations
FulcrumSec, a data extortion group active since September 2025, exploits hardcoded credentials, unpatched applications, and misconfigured storage to breach organizations and steal data. Rather than using artificial intelligence (AI) for initial compromise, the group leverages AI during ransom negotiations to generate pressure on victims for higher payouts. The group reports breaching 25 organizations and exfiltrating several terabytes of data.
Why it matters: Organizations targeted by extortion groups face increased ransom demands when attackers use AI-generated leverage in negotiations; practitioners should prioritize eliminating hardcoded credentials, patching vulnerable applications, and auditing storage configurations to prevent the initial access vectors this group exploits.
- threat intel
Police Disrupt a €140M Cyber Fraud Ring in Spain
Spanish police dismantled a cybercriminal network that conducted multiple cyberattacks and laundered approximately 140 million euros through intricate financial schemes. The operation targeted the organized fraud ring's infrastructure and money laundering operations.
Why it matters: Organizations across Europe should review their fraud detection and financial monitoring controls, as this ring's sophistication in attack variety and money laundering indicates ongoing threats to financial institutions and their customers.
- vulnerabilities
Nightmare Eclipse Drops ‘LegacyHive’ Windows Zero-Day
A researcher has disclosed a Windows zero-day vulnerability named LegacyHive, attributed to the threat actor Nightmare Eclipse. The researcher released a stripped proof-of-concept to limit immediate weaponization of the flaw.
Why it matters: Windows administrators and security teams need to monitor for patches and defensive guidance from Microsoft, as this zero-day presents an active threat to unpatched systems until remediation is available.
- vulnerabilities
Trend Micro, Tanium, ESET and Tenable Patch Severe Product Vulnerabilities
Cybersecurity vendors Trend Micro, Tanium, ESET, and Tenable released patches for critical and high-severity vulnerabilities in their products. The article provides no details on affected products, vulnerability types, or patch availability.
Why it matters: Security teams running these tools should immediately check vendor advisories to determine if patches apply to their deployments and prioritize installation based on severity ratings.
- regulatory
Reading between the lines of a cyber insurance policy
The global cyber insurance market reached approximately $16 billion in premiums in 2024, with coverage now widespread across regulated industries, though payouts have become less predictable. The Global Federation of Insurance Associations identified a significant protection gap of roughly $900 billion, indicating a substantial mismatch between actual cyber exposures and available coverage.
Why it matters: Risk managers and security leaders in regulated industries need to understand the true limits of their cyber insurance coverage and identify the exposures that remain unprotected by their policies.
- research
What public money does to open-source projects
Open-source software comprises approximately 96 percent of enterprise codebases, yet most of it is maintained by unpaid volunteers. Recent supply chain incidents like the log4j vulnerability in December 2021 and the xz utils backdoor in 2024 have highlighted the risks of relying on under-resourced projects. The article examines how public funding affects open-source project sustainability and security.
Why it matters: Engineering and security teams depend on open-source libraries for critical infrastructure and should understand how funding models and maintainer capacity directly influence vulnerability detection and patch velocity.
- vulnerabilities
Companies keep getting breached by vulnerabilities they already knew about
A survey of 300 IT and cybersecurity leaders in the United States and United Kingdom found a significant gap between vulnerability discovery and remediation. Organizations identify more weaknesses than ever before, but struggle with the processes of assigning, approving, deploying, and confirming fixes, leaving known vulnerabilities exposed to breach risk.
Why it matters: Security teams and IT leaders face a gap in vulnerability management workflows that leaves discovered weaknesses unpatched, creating exploitable gaps that attackers actively target.
- ai security
GPT-Red beat human red teamers on a prompt injection test
OpenAI has developed GPT-Red, an automated red-teaming model trained through self-play reinforcement learning to identify prompt injection vulnerabilities in large language models. In testing, GPT-Red outperformed human red teamers at finding weaknesses by iterating through prompts to achieve objectives such as data exfiltration, with both attacker and defender models learning simultaneously across multiple scenarios.
Why it matters: Security teams and LLM builders need to understand that automated adversarial techniques can discover prompt injection flaws faster than manual testing, informing investment in both defensive mechanisms and adversarial testing practices.
- threat intel
Finance phishing works because it sounds boringly normal
Finance departments receive high volumes of routine business emails, making them a target for phishing campaigns that blend in as legitimate correspondence rather than using obvious urgency tactics. Cofense research shows that attackers craft these emails to resemble normal invoices, contracts, and payment notices, which can evade artificial intelligence (AI)-based email security gateways and other defenses. Threat actors deliberately leverage the predictability of financial workflows to increase the likelihood of successful compromise.
Why it matters: Finance and accounts payable teams face elevated phishing risk because credential theft or payment diversion attacks use legitimate-sounding messages that bypass both technical controls and human vigilance; practitioners should audit email security rules and user training to focus on workflow-based anomalies, not just sender reputation.
- threat intel
AI Has Enhanced Iran’s Asymmetric Playbook During the 2026 Conflict
Between January and June 2026, Iran leveraged artificial intelligence to enhance its asymmetric warfare capabilities across cyber operations, information warfare, propaganda production, and domestic surveillance during military and political crises. AI functioned as a force multiplier that increased the speed, scale, and effectiveness of Iranian operations, particularly in information campaigns and cyber attacks, though its direct impact on battlefield tactics remains unconfirmed. Iran's hybrid approach, augmented by partnerships with Russia and China for military and surveillance technologies, demonstrates how AI amplifies existing capabilities rather than creating fundamentally new ones.
Why it matters: Security practitioners and critical infrastructure operators in Western and regional organizations face elevated threats from AI-enhanced Iranian cyber operations, drone attacks, and influence campaigns targeting both digital assets and physical infrastructure, necessitating improved defenses against attribution-resistant asymmetric threats.
- breaches incidents
Cyberattack on Japan's largest cold-chain operator disrupts KFC, supermarket supplies
A cyberattack on Nichirei Logistics Group, Japan's largest cold-chain operator, disrupted ingredient deliveries to Kentucky Fried Chicken and other major restaurant chains. The attack caused operational strain for food service businesses dependent on the logistics provider's supply capabilities.
Why it matters: Restaurants and food service operators using Nichirei face immediate supply chain disruptions; practitioners should assess critical dependencies on third-party logistics providers and develop contingency plans for supplier outages.