TeamViewer on DanSec Daily

Cybersecurity Brief – 2026-10-31

Sat, 31 Oct 2026 00:00:00 +0000

Major Incidents or Breaches

Chainalysis suffered a cybersecurity incident in which an external actor accessed internal tools and non-sensitive customer data, including customer names, email addresses, and user IDs. The attack reportedly leveraged “offensive agentic techniques”^[1]^[9].
Fashion retailer Resolve disclosed a breach resulting in the theft of customer payment card information^[27].
The Florida Department of Education revealed that 4.9 million individuals had their Social Security numbers leaked following a breach of the Florida Student Financial Aid website^[28].
Trezor warned of a suspected hardware implant in a Trezor Safe 5 wallet sold by a reseller in Southeast Asia, potentially enabling cryptocurrency theft^[11].
A malicious version of Anthropic’s Claude Code extension was distributed via the Cursor marketplace for approximately 11 hours, exposing up to 5,000 developers to malware that planted persistent SSH backdoors^[12]^[18]^[23]^[32]^[47].
Apple notified users in 99 countries of mercenary spyware infections, ahead of international negotiations on guidelines for the use of commercial hacking tools^[35].

Newly Discovered Vulnerabilities

Attackers are chaining three Ivanti Connect Secure vulnerabilities to gain unauthenticated root access. Two of these flaws were recently added to CISA’s Known Exploited Vulnerabilities catalog^[2].
A critical Linux kernel vulnerability affecting MPTCP servers allows unauthenticated remote attackers to execute commands as root^[10].
TeamViewer patched a vulnerability where a remote computer could attack the computer controlling it during a session^[6].
A high-severity remote code execution flaw in WordPress Core, not yet assigned a CVE, has been targeted in nearly 63,000 attack attempts within 24 hours of patch release^[7].
An origin-validation flaw in BeyondTrust Remote Support (CVE-2026-102764) allows unauthenticated attackers to upload files and achieve remote code execution as the appliance OS user^[49].
Over 22.6 million exposed instances are potentially affected by critical PHP flaws exploited in PHP2Shell attacks, which have successfully compromised Palo Alto firewalls but not WordPress^[15]^[41].
Flaws in SS7 and Diameter signaling protocols allow attackers to track phone locations, intercept communications, and impersonate subscribers without compromising the device itself^[13].

Notable Threat Actor Activity

Chinese state-sponsored group Vapor Panda used the AI-powered, Go-based ARTEX penetration testing framework to compromise over 200 organizations in at least 14 countries, with 140 sensitive targets breached in 11 days^[5]^[34].
Syrian hacker Ahmad Umar Agha, associated with the Syrian Electronic Army and pro-Assad cyber operations, was sentenced in the US to 42 months in prison for attacks on Western media, businesses, and government entities^[14]^[29]^[38].
Threat actors launched approximately 8,226 network-layer DDoS attacks per hour in Q3, totalling 18.2 million attacks, representing a 38 percent increase according to Cloudflare^[4].
A new infostealer malware dubbed Volta-Autodoxer uses AI to collect credentials, credit card data, sessions, and other sensitive information, and leverages LLMs to profile and prioritize victims’ assets^[22]^[45].
Romance scammers are targeting US citizens to act as money mules, leveraging social engineering techniques to facilitate laundering operations^[39].
A criminal group spent over $7,000 on AI services to enhance extortion attacks, leaving digital evidence of their activities^[30].

Trends, Tools, or Tactics of Interest

AI-powered penetration testing frameworks, such as ARTEX, are being adopted by state-sponsored actors for rapid and scalable compromise of high-value targets^[5]^[34].
The use of large language models (LLMs) by infostealer malware, as demonstrated by Volta-Autodoxer, is enabling automated victim profiling and prioritization of stolen data^[22]^[45].
The supply chain attack on the Claude Code extension for Cursor highlights the risks of malicious code distribution via third-party marketplaces, with attackers using persistent SSH backdoors to maintain access to developer systems^[12]^[18]^[23]^[47].
Fake Chrome extensions, including PDF and search tools, hijacked searches for 13.4 million users by routing queries through attacker-controlled servers^[46].
Abandoned or “undead” software projects continue to pose security risks due to unmaintained infrastructure and vulnerable domains^[21].
Attackers are exploiting help desk impersonation and abusing Active Directory Certificate Services (AD CS) through authentication relay and Kerberos attacks to escalate privileges^[48].
Passkeys are increasingly adopted but attackers are shifting tactics to target areas around authentication rather than the authenticator itself^[17].

Regulatory or Policy Developments Affecting the Security Industry

The US Department of Justice raised privacy concerns regarding the potential sale of Character.AI chat data, recommending the appointment of a consumer privacy ombudsman to advise on data handling and sale terms^[26].
The Pall Mall Process, an international negotiation on guidelines for the use of commercial hacking tools, is nearing finalization as Apple issues spyware notifications globally^[35].
General Motors agreed to a $220 million settlement regarding multistate allegations of improper driver tracking^[31].
A new framework, RAID-AI, has been introduced to help organizations assess and improve their resilience to ransomware attacks targeting AI systems^[33].

Sources

Chainalysis Confirms Intrusion Into Internal Tools and Exposure of Customer Identity Data - thehackernews.com
Chainalysis discloses data breach from "offensive agentic techniques" - bleepingcomputer.com
Fashion retailer Resolve says hackers stole payment card information - therecord.media
Florida warns that 4.9 million people had Social Security numbers leaked in financial aid site breach - therecord.media
Trezor warns of suspected crypto-theft implant in hardware wallet - bleepingcomputer.com
Malicious Claude Code extension on Cursor planted SSH backdoors - bleepingcomputer.com
Devs Hit by Trojanized Claude Code Extension for Cursor - darkreading.com
A Supply Chain Attack Put Malicious Anthropic Claude Extensions on Thousands of Devices - wired.com
Anthropic: That Claude Code extension compromise… couldn't be us - theregister.com
Inside a malicious Claude Code-branded extension and its persistent SSH backdoor - microsoft.com
Apple notifications alert victims in 99 countries about mercenary spyware infections - cyberscoop.com
Attackers Chain Three Ivanti Connect Secure Flaws to Gain Unauthenticated Root Access - thehackernews.com
Critical Linux kernel flaw allows remote root attacks on MPTCP servers - bleepingcomputer.com
TeamViewer Flaw Lets Remote Computers Attack Users Who Connect to Control Them - thehackernews.com
WordPress RCE Flaw Draws Nearly 63,000 Attack Attempts in Just 24 Hours - thehackernews.com
Rapid7 Analysis: Unauthenticated File Upload to RCE in BeyondTrust Remote Support (CVE-2026-102764) - rapid7.com
PHP2Shell Attacks Root Palo Alto Firewalls but Fail Against WordPress - securityweek.com
This month in security with Tony Anscombe – October 2026 edition - welivesecurity.com
SS7 and Diameter Attacks: How Signaling Flaws Expose Your Phone - bleepingcomputer.com
Chinese Hackers Use AI-Powered ARTEX to Target Over 200 Organizations Worldwide - thehackernews.com
China-backed threat group used agentic AI pentesting tool to compromise 140 sensitive targets in 11 days - cyberscoop.com
Syrian Hacker Gets Prison Time in US After Over a Decade on the Run - securityweek.com
Syrian Electronic Army hacker gets prison time for attacks on Western targets - therecord.media
Syrian hacker who served Assad regime gets 42 months in US prison - cyberscoop.com
Cloudflare Reports 18.2 Million Network-Layer DDoS Attacks in Q3, Up 38% - thehackernews.com
Spooky New Infostealer Malware Uses AI to ‘Autodox’ Victims - wired.com
Appraisal by AI: Model-driven victim prioritization in the VOLTA infostealer - blog.talosintelligence.com
Warning: Romance Scammers Are Targeting US Citizens For Money Laundering - blog.knowbe4.com
Crim spent $7K+ on AI services in extortion attacks, then left evidence for all to see - theregister.com
Fake PDF and search tools hijacked searches for 13.4 million Chrome users - malwarebytes.com
Undead Software Projects Continue to Raise Security Risks - darkreading.com
From help desk impersonation to AD CS abuse: Investigating ESC1 and ESC8 intrusion paths - microsoft.com
Passkeys Are Better Than Passwords. But They Are Not Enough. - securityweek.com
DOJ raises privacy concerns over potential sale of Character.AI chats - therecord.media
GM cracks the petty cash box for multistate driver tracking settlement - theregister.com
New Framework Puts Ransomware Resilience for AI Systems in Focus - infosecurity-magazine.com

Cybersecurity Brief – 2026-10-30

Fri, 30 Oct 2026 00:00:00 +0000

Major Incidents or Breaches

Binance, the world’s largest cryptocurrency exchange, suffered a record-setting security breach resulting in the theft of approximately $6.3 billion in digital assets. The incident has been linked to the compromise of a DevOps engineer’s laptop, and blockchain investigators attribute the attack to North Korean threat actors, specifically the TraderTraitor group. Nearly $3 billion in assets were reportedly frozen in response to the breach, which was executed in under 20 minutes ^[7]^[15]^[17]^[32]^[42]^[47]^[49]^[54]^[61]^[79].
Polymarket, a prediction market platform, temporarily halted trading after hackers compromised a third-party service and distributed a cryptocurrency-draining JavaScript, resulting in the theft of approximately $2.5 million from user accounts. The attack leveraged a supply chain vector via a website monitoring tool ^[20]^[29]^[45].
A searchable database containing records of more than 41,000 current and former US Department of Homeland Security employees and contractors was leaked, exposing contact details and sensitive job information ^[66]^[74]^[80].
Belgium’s tax agency confirmed a cyberattack that forced it to pull services offline and extend tax deadlines ^[52].
Japan Airlines disclosed a third-party data breach at Amadeus’s Passenger Service System, exposing data of about 92,000 customers ^[19].
An employee at RIZAP Inc. in Japan uploaded health records of 726 individuals to a personal AI chat, resulting in a sensitive data exposure incident ^[71].
The University of New Mexico Health reported a cyberattack, with a ransomware group claiming responsibility ^[69].
The Florida Department of Education revealed a breach of its student financial aid website, potentially exposing personal information, including Social Security numbers, of approximately 4.9 million people ^[70].
Frontline school software suffered a breach exposing student Social Security numbers and Medicaid IDs ^[73].

Newly Discovered Vulnerabilities

A highly critical authorization bypass vulnerability in Drupal Commerce Core, tracked as CVE-2026-118539, is being actively exploited in the wild to target online stores ^[5]^[14]^[28].
A maximum-severity remote code execution vulnerability in GitLab, CVE-2026-97229, is now being exploited against unpatched servers, including Alpine Linux deployments ^[6]^[12]^[30].
The Linux kernel is affected by a critical TCP flaw, CVE-2026-98432, rated 9.8 CVSS, impacting major distributions such as Ubuntu and Red Hat Enterprise Linux ^[9].
Cisco Catalyst SD-WAN Controllers are being targeted via CVE-2026-20127, an authentication-bypass vulnerability exploited to deploy custom KeyDrift malware ^[6]^[22].
TeamViewer is affected by an integer overflow vulnerability, CVE-2026-111585, which could allow malicious hosts to execute arbitrary code on support technicians’ systems ^[31].
Researchers found critical vulnerabilities in five AI developer tools, with flaws in OpenAI and Microsoft products still unpatched, primarily involving embedded browsers ^[50].
Critical vulnerabilities in PHP, CVE-2026-109618 and CVE-2026-109619, are being targeted for remote code execution in web server environments ^[81]^[77].
A fresh unauthenticated race condition vulnerability in WordPress is under active exploitation, allowing remote code execution ^[27].

Notable Threat Actor Activity

North Korean threat actors, specifically the TraderTraitor group, are linked to the $6.3 billion Binance heist. Microsoft also reports increased adoption of AI tools by North Korean actors to enhance attack speed and efficiency ^[32]^[42]^[54]^[79]^[82].
A suspected Chinese-nexus actor compromised a US government-affiliated network by chaining a custom PHP2Shell exploit to gain root access to Palo Alto Networks firewall servers ^[51]^[53]^[67].
UAT-11473, a newly tracked APT group, exploited a Cisco SD-WAN flaw to deploy custom KeyDrift malware ^[22].
Belarusian state hackers, identified as the “FrostyNeighbor” APT, launched a destructive data-wiping attack in Ukraine, aligning with Russian cyber objectives ^[36]^[67].
The Syrian Electronic Army hacker “Th3Pr0” was sentenced to 42 months in US federal prison for a hacking conspiracy that involved posting a fake tweet, triggering a $136 billion market drop ^[3]^[56].
Threat actors are distributing malicious loaders and cryptocurrency miners disguised as AI desktop applications for Anthropic Claude and OpenAI ChatGPT, abusing signed executables ^[10].
A trojanized version of Anthropic’s Claude Code extension was distributed via the Cursor Marketplace, planting a persistent SSH backdoor on developer machines ^[4]^[59].
A campaign abusing Hugging Face models for crypto mining has resurfaced, hijacking GPU servers and deploying backdoors ^[18].
Attackers exploited a fake cryptocurrency trading app on the Apple App Store, resulting in the theft of $23,000 from 27 Solana wallets ^[21].
A bulletproof hosting provider, Virtualine, lost its domain to an FBI seizure after supporting ransomware and infostealer operators ^[26].

Trends, Tools, or Tactics of Interest

Attackers are increasingly leveraging supply chain attacks, as seen in the Polymarket incident and the trojanized Claude Code extension ^[20]^[29]^[4]^[59].
There is a growing trend of threat actors exploiting vulnerabilities in widely used open-source and enterprise platforms, including Drupal, GitLab, Linux, PHP, WordPress, and Cisco SD-WAN ^[5]^[6]^[9]^[12]^[14]^[22]^[27]^[28]^[30]^[31]^[50]^[77]^[81].
AI is being adopted both as an attack vector and as a defensive tool. North Korean threat actors are integrating AI into their tradecraft, while Google and Anthropic highlight AI models capable of discovering and exploiting vulnerabilities. AI-assisted remediation and agent verification tools are being launched for enterprise security ^[34]^[40]^[57]^[58]^[60]^[82].
Phishing campaigns themed around Anthropic are bypassing leading secure email gateway solutions ^[37].
Attackers are abusing legitimate communication channels, such as Microsoft Teams and phone calls, to conduct vishing and social engineering attacks that result in remote access session compromises ^[62].
Chrome extensions with over 13 million installs were found hijacking search queries for ad revenue ^[23].
Smart TVs and IoT devices continue to present significant security risks due to unpatched vulnerabilities and data tracking ^[65].
Criminal groups rapidly rebuild fake shopping sites after takedowns, often reusing web assets and infrastructure ^[64].
Malware operators are using victims’ OpenAI API keys to execute AI queries via NIK C2 malware ^[55].

Regulatory or Policy Developments Affecting the Security Industry

The UK’s Ofcom has outlined four possible age-verification methods for enforcing the under-16 social media ban, including photo ID, bank checks, and face scans ^[78].
The UK Investigatory Powers Commissioner’s Office reported that GCHQ cannot account for thousands of data storage devices, expressing concern over the agency’s asset management and related security implications ^[43]^[48].
Australia’s Signals Directorate disrupted the Amadey malware operation, severing criminal access to approximately 180,000 infected computers ^[72].

Sources

Binance Suffers Security Breach Affecting $6.3 Billion in Assets, Nearly $3 Billion Frozen - thehackernews.com
Binance links $6.3 billion crypto theft to compromised DevOps laptop - bleepingcomputer.com
Binance probes potential $5.8 billion crypto-theft hack - bleepingcomputer.com
$6 Billion Binance Hack Bears Hallmarks of North Korea’s TraderTraitor - securityweek.com
The $6.3 Billion Binance Hack Was Pulled Off in Just 20 Minutes - wired.com
Binance hit by record $6.3 billion crypto hack, scrambles to freeze funds - therecord.media
In the year AI re-wrote the infosec rulebook, it’s comforting to know crypto exchanges still have bad security - theregister.com
Binance Hit by Record-Breaking $6.3bn Hack in Largest Cryptocurrency Heist in History - infosecurity-magazine.com
Binance hit by $6.3 billion theft in what could be the biggest crypto heist on record - helpnetsecurity.com
Binance hit by largest-ever crypto theft, with $6.3 billion stolen - malwarebytes.com
Polymarket halts trading after crypto-draining script breach, $2.5M traced - bleepingcomputer.com
$2 Million+ Drained From Polymarket Traders in New Supply Chain Attack - securityweek.com
Polymarket confirms cyberattack on third-party provider after outages - therecord.media
Leaked DHS employee archive, Icepick, faces challenges to stay online - cyberscoop.com
DHS personnel leak exposes contact details and sensitive job information - databreaches.net
Hackers steal and publish data on 41,000 Homeland Security workers - malwarebytes.com
Belgium's tax agency confirms cyberattack, pulls services offline - theregister.com
Japan Airlines suffers third-party data breach exposing passenger info - bleepingcomputer.com
A RIZAP Employee Uploaded 726 People’s Health Records to a Personal AI Chat - databreaches.net
UNM Health addresses cyberattack in New Mexico after ransomware group claims responsibility - databreaches.net
Florida education department says data breach may affect 4.9 million people - databreaches.net
Student SSNs and Medicaid IDs Exposed in Frontline School Software Breach - databreaches.net
Hackers Actively Exploit Highly Critical Drupal Commerce Core Authorization Bypass Flaw - thehackernews.com
Drupal warns of critical Commerce Core flaw exploited in attacks - bleepingcomputer.com
Drupal Commerce Core Vulnerability Exploited Against Online Stores - securityweek.com
CISA Adds Actively Exploited GitLab RCE and Cisco Firewall DoS Flaws to KEV - thehackernews.com
CISA says max severity GitLab flaw now exploited in the wild - bleepingcomputer.com
Max-Severity GitLab Vulnerability Exploited Against Alpine Linux - securityweek.com
Linux Kernel TCP Flaw With 9.8 CVSS Score Affects Ubuntu, Red Hat and Other Distributions - thehackernews.com
Hackers exploit Cisco SD-WAN flaw to deploy custom KeyDrift malware - bleepingcomputer.com
TeamViewer Flaw Could Let Malicious Hosts Take Over Support Technicians’ Machines - securityweek.com
Researchers find bugs in 5 AI developer tools, with OpenAI and Microsoft flaws still unpatched - theregister.com
Threat Brief: Critical Vulnerabilities in PHP: CVE-2026-109618, CVE-2026-109619 (Updated October 29) - unit42.paloaltonetworks.com
A quiet week in infosec <.< - blog.talosintelligence.com
Fresh WordPress Vulnerability in Attackers’ Crosshairs - securityweek.com
Tracking AI adoption in North Korean threat actor tradecraft - microsoft.com
Suspected Chinese attackers built a custom PHP2Shell exploit and rooted 25 servers - theregister.com
Chinese Actor Compromises US Government Network Via Chained PHP2Shell Exploit - infosecurity-magazine.com
Risky Bulletin: Belarusian state hackers get on the data-wiping train - news.risky.biz
As Russia's Cyberwar Evolves, Belarus Steps Forward - darkreading.com
Syrian Hacker Behind Fake Tweet That Triggered $136 Billion Market Drop Gets 42 Months in Prison - thehackernews.com
Assad regime hacker who caused $136 billion market drop sentenced to 42 months - helpnetsecurity.com
Fake ChatGPT and Claude Apps Abuse Signed Executables to Deploy Crypto Miners - thehackernews.com
Official Claude Code Extension on Cursor Marketplace Trojanized with Persistent SSH Backdoor - thehackernews.com
Claude Code extension poisoned with persistent SSH backdoor - helpnetsecurity.com
Hugging Face malware campaign returns to hijack GPUs for crypto mining - bleepingcomputer.com
Fake crypto app on Apple App Store linked to $23K wallet theft - bleepingcomputer.com
Bulletproof Host Virtualine Loses Its Domain to FBI Seizure - securityweek.com
Google Touts Powerful Cybersecurity Capabilities in Gemini 4 Carbon Model - securityweek.com
Claude Model Builds Exploits Across Major Browsers & OSes - darkreading.com
Shufti ties AI agent actions to verified owners and permissions - helpnetsecurity.com
Bedrock Data routes data exposure fixes to owners and verifies results - helpnetsecurity.com
Open-weight AI model GLM-5.3 cheated its way to correct answers - helpnetsecurity.com
Anthropic-Themed Phishing Emails Bypass Top Security Tools - darkreading.com
Fake IT help desk callers got by on little more than names and email addresses - helpnetsecurity.com
Chrome extensions with 13.4 million installs caught hijacking searches - bleepingcomputer.com
Smart TVs can harbor unpatched security flaws and track what you watch - helpnetsecurity.com
Scammers rebuild fake shopping sites within hours of a takedown - helpnetsecurity.com
NIK C2 Malware Can Use Victims’ OpenAI Keys for AI Queries - infosecurity-magazine.com
Photo ID, bank checks, face scan? The UK’s age-check options - malwarebytes.com
British spy agency cannot account for thousands of data storage devices, watchdog says - therecord.media
British intelligence agency misplaced 'thousands' of storage devices, years later still can't say how many - theregister.com
Australia severs criminal access to 180,000 computers after breaching Amadey malware backend - databreaches.net

Cybersecurity Brief – 2026-09-30

Wed, 30 Sep 2026 00:00:00 +0000

Major Incidents or Breaches

Cryptocurrency exchange Bitget suffered a breach resulting in a $387.5 million loss after attackers exploited a zero-day vulnerability in a third-party security product used by the company ^[15].
The French tax administration experienced a data breach in which an attacker used stolen staff credentials to exfiltrate tax data belonging to hundreds of thousands of individuals and businesses. The compromise went undetected for seven weeks ^[9].
The Dutch Institute for Vulnerability Disclosure (DIVD) was breached via an automated AI agent in an attack described as “loud and very, very messy” ^[24].
South Africa’s air traffic control systems were targeted by a cyberattack that deployed a ransomware toolkit on at least one operational network ^[25].
AI coding agents inadvertently exposed more than 13,000 internal company images, including billing records, by uploading screenshots to public GitHub repositories ^[4].

Newly Discovered Vulnerabilities

Multiple critical vulnerabilities have been disclosed and patched in Citrix NetScaler ADC and Gateway products, including CVE-2026-88772 and CVE-2026-88771. These flaws allow pre-authentication code execution and have been actively exploited to gain root access, deploy web shells, steal credentials, and move laterally within victim networks. Government and finance organizations are among those affected ^[6]^[8]^[20]^[30]^[36].
OpenSSL released fixes for a high-severity DTLS flaw that can leak heap memory to the other side of a DTLS connection or cause a crash. WolfSSL also patched several high-severity vulnerabilities ^[7]^[41].
TeamViewer warned users to urgently patch a set of high-severity vulnerabilities in its client and host software ^[14].
A newly disclosed Spectre-v2 Branch Target Reuse (BTR) attack variant enables attackers to recover Linux root password hashes in minutes, bypassing existing mitigations ^[10]^[23].
Apple reported that CVE-2026-86950, an out-of-bounds write vulnerability, is being actively weaponized in targeted attacks ^[26].
WatchGuard issued patches for 15 vulnerabilities in Fireware OS, including code execution, denial of service, authorization, and path traversal bugs ^[35].
Unsloth Studio patched a vulnerability allowing malicious AI models to execute arbitrary Python code during model inspection if the trust_remote_code setting is enabled ^[27].
Chrome and Firefox released updates addressing over 100 vulnerabilities, some of which could allow remote code execution or sandbox escape ^[37].

Notable Threat Actor Activity

Russian state-sponsored group Star Blizzard (also known as APT29) has conducted large-scale phishing campaigns using fake event invitations and the ‘RedFlick’ infection chain to deliver the CosmicPulse backdoor, targeting over 100 organizations ^[11]^[39].
Microsoft observed a China-based actor, dubbed ‘NeedyMantis’, using a previously unidentified malware framework for long-term access in targeted intrusions against telecommunications, universities, medical, and government-related organizations ^[29].
The ShinyHunters extortion group was addressed by the FBI following the arrest of an alleged leader, with the FBI urging remaining members to turn themselves in. ShinyHunters responded defiantly ^[19]^[40].

Trends, Tools, or Tactics of Interest

Social engineering attacks are increasingly leveraging audio and video deepfakes, with 41 percent of CISOs reporting an audio deepfake attack against their organization in the past year and 36 percent reporting deepfake targeting of employees ^[1].
Custom malicious ChatGPT variants promoted via sponsored Google results are being used to lure victims to sites that deploy RAT malware through ClickFix attacks ^[18].
Recent US-focused phishing campaigns targeting C-suite executives have been observed stealing Microsoft 365 sessions and deploying remote monitoring and management (RMM) tools for persistent remote access. The campaigns predominantly targeted technology, manufacturing, government, and consulting sectors ^[5].
Browser-based attacks remain a significant initial vector for breaches, with many attacks contained entirely within browser sessions ^[3].
AI agent security is being addressed with new tools such as KnowBe4’s Agent Risk Manager, which integrates with the Claude Compliance API to monitor and detect AI agent risks in real time ^[2].
Anthropic and other industry voices are raising concerns over liability risks associated with autonomous AI agents, as lawsuits emerge following incidents involving AI-driven attacks ^[33]^[38].

Regulatory or Policy Developments Affecting the Security Industry

Microsoft will implement new protections in Entra ID to block external script injection attacks starting in October ^[13].
Cloudflare announced a public certificate authority designed for post-quantum cryptography, offering automated certificates intended to be resistant to quantum computing threats ^[28].
A group of major technology firms has signed a voluntary accord to self-police AI development, focusing on four key steps and signaling the possibility of future regulation ^[42].
DARPA selected Xint to use AI for identifying vulnerabilities in military messaging apps, with potential commercial applications ^[44].

Sources

Bitget hacked via zero-day in third-party security products - bleepingcomputer.com
French Tax Data Theft Using Stolen Staff Passwords Went Undetected for Seven Weeks - thehackernews.com
Automated AI agent used to breach cybersecurity nonprofit DIVD - bleepingcomputer.com
South Africa Seeks Help After Cyberattack Targets Air Traffic Control - darkreading.com
AI Coding Agents Exposed 13,000 Internal Images, Including Billing Records, on GitHub - thehackernews.com
Attackers Exploit NetScaler Flaw for Root Access, Deploy WHIPSHOT and SLAPSHOT - thehackernews.com
Citrix NetScaler CVE-2026-88772 Exploit Details Show Pre-Auth Path to Shellcode Execution - thehackernews.com
Hackers exploit Citrix NetScaler zero-day to deploy web shells - bleepingcomputer.com
Dual NetScaler Zero-Days Trigger Chaos for Citrix Customers - darkreading.com
Government, Finance Orgs Targeted in Weeks-Long NetScaler Zero-Day Attacks - securityweek.com
OpenSSL Fixes High-Severity DTLS Flaw That Can Leak Heap Memory Unencrypted - thehackernews.com
High-Severity Vulnerabilities Patched in OpenSSL, WolfSSL - securityweek.com
TeamViewer urges users to patch severe flaws “as soon as possible” - bleepingcomputer.com
New Spectre-v2 BTR Attack Leaks Linux Memory Despite Existing Defenses - thehackernews.com
New Spectre v2 attack variant leaks Linux root password hash in minutes - bleepingcomputer.com
Apple Zero-Day Vulnerability Weaponized in Targeted Attacks - darkreading.com
WatchGuard Patches Critical Fireware OS Code Injection Vulnerability - securityweek.com
Unsloth Studio Flaw Turns Routine Model Inspection Into Code Execution - darkreading.com
Chrome, Firefox Updates Patch Over 100 Vulnerabilities - securityweek.com
Russia's Star Blizzard Targets 100+ Organizations With Fake Event Invites to Deliver Backdoor - thehackernews.com
Russian APT Star Blizzard Uses ‘RedFlick’ Infection Chain in Recent Attacks - securityweek.com
'NeedyMantis' Provides Long-Term Access to Compromised Networks - darkreading.com
FBI tells ShinyHunters members to turn themselves in after recent arrest - bleepingcomputer.com
ShinyHunters Defiant After FBI Calls on Members to Come Forward - securityweek.com
Report: Social Engineering Attacks Are Increasingly Using Audio and Video Deepfakes - blog.knowbe4.com
Custom ChatGPTs push ClickFix attacks to deploy RAT malware - bleepingcomputer.com
US-Focused CSuite Phishing Steals Microsoft 365 Sessions and Deploys RMM Tools for Remote Access - thehackernews.com
Know Your Enemy: Browser-Based Attack Techniques in 2026 - thehackernews.com
Govern Claude AI with KnowBe4’s Agent Risk Manager - blog.knowbe4.com
OpenAI Gets Sued Over the Hugging Face Hack - wired.com
Anthropic Flags AI Agent Liability Risks as OpenAI Faces Hacking Lawsuit - securityweek.com
Microsoft to block Entra ID script injection attacks starting October - bleepingcomputer.com
Cloudflare Announces Public Certificate Authority for the Post-Quantum Web - darkreading.com
Trump Says Top Tech Firms Have Signed Accord to ‘Self-Police’ AI Development - securityweek.com
DARPA Selects Xint to Use AI in Securing Military Messaging Apps - securityweek.com