Every notable cyber threat for education and government (125 items), including CVE references.
AI-assisted summaries and fixes, sources cited.
High · Help Net Security ·
An unidentified intruder reached a refrigeration vendor used by three blood banks, changed cold-room parameters and caused a warming event that was detected and reversed.
Medium · Hackread ·
Intruders changed cooling parameters in grocery refrigeration systems, disrupting European stores and UK supply partners and leading to spoiled produce and discarded food.
High · The Hacker News ·
Rails maintainers disclosed a high-severity Active Storage issue in which unauthenticated crafted uploads processed through ImageMagick can write files to the server.
High · The Hacker News ·
Xen Project published fixes on November 3 for a hypervisor flaw that lets a malicious guest read and write memory of co-resident VMs and dom0; no workaround is known.
High · Help Net Security ·
Unknown attackers hijacked OPSShield's cPGuard update path and used fake releases to compromise almost 3,000 Linux hosting servers, exposing hashed passwords and contact data.
High · SecurityWeek ·
Attackers used stolen login credentials to reach OpsShield's release infrastructure and serve tainted cPGuard builds across several versions; no CVE has been assigned.
High · SecurityWeek ·
Unknown attackers allegedly compromised three Binance engineers through a booby-trapped coding task and misuse of an AI tool, leading to stolen records and a $6 billion cryptocurrency theft.
Medium · BleepingComputer ·
A broad Russian-origin operation is using hijacked websites and counterfeit portals to deliver no-interaction exploits against iPhone and Android users, risking handset takeover.
High · BleepingComputer ·
Attackers reportedly compromised cPGuard's automatic update path to push a backdoor and a data-stealing script to users of the cPGuard suite.
High · BleepingComputer ·
An intrusion into a WorkOS internal sales application tied to ARTEX AI exposed business contact records and summaries of customer conversations.
Medium · BleepingComputer ·
Attackers are abusing cloud AI services and large language models, generating more than $100,000 in fraudulent charges per day and a seven-figure total, with no CVE to patch.
High · The Hacker News ·
mSeven Software alerted mSecure users on November 3 to reset credentials, re-enroll 2FA, and remove passkeys after detecting suspicious account activity.
Medium · The Record ·
Unknown intruders manipulated refrigeration controllers, causing defrosting and spoiled inventory at supermarkets in Europe and Britain during a second wave of disruption.
Medium · The Hacker News ·
Adversaries are abusing weak configurations to take over Google-hosted model access, steal tokens, and run up cloud charges approaching seven figures at affected organizations.
High · Dark Reading ·
Cybercriminal crews are hijacking cloud accounts to run and resell large language model capacity, leaving victims with unauthorized charges and revoked access.
High · The Hacker News ·
An unidentified party tampered with OPSSHIELD's cPGuard update path on November 2, distributing a malicious script that opened backdoors and stole credentials and files from customer Linux hosts.
High · BleepingComputer ·
Barracuda is advising customers to update its appliances after two vulnerabilities were identified that could grant attackers root privileges and the ability to run commands.
High · The Record ·
A Russia-tied espionage operation is using lightweight malicious code to lift documents and login data from Ukrainian groups tied to war-crimes probes and military supply routes.
Medium · SecurityWeek ·
Repeated exploitation attempts against Node-RED installations have been traced to one IP address, with attackers delivering Linux bot and webshell payloads.
High · Hackread ·
Malicious add-ons for Cursor and Anthropic’s Claude Code have been linked to hidden SSH access, persistence, and device profiling on developer machines.
High · The Hacker News ·
A Russia-aligned threat group used a WinRAR weakness to deliver data-stealing malware to Ukrainian government, justice, law enforcement, and defense organizations in February 2025.
High · Help Net Security ·
Google's threat team and Mandiant document multiple 2026 intrusions where attackers seized cloud credentials, launched AI workloads, and drove daily costs past $100,000.
Medium · SecurityWeek ·
NATO's cyber defence centre of excellence used OpenAI's Daybreak to bring AI agents into a joint cyber-kinetic training exercise hosted in Estonia.
High · The Hacker News ·
Unidentified actors abused a weakness in BTCPay Server releases up to 2.4.1 to compromise hot wallets and steal Bitcoin and Litecoin, prompting urgent upgrades.