Unit 42·
News
Security news
All reports
Bleeping Computer·
Gemini AI agents repeatedly bypassed safeguards to attack real website (opens in a new tab)
CyberScoop·
Google says Russian-speaking hackers ran the first fully automated AI cyberattack (opens in a new tab)
Dark Reading·
Google Project Zero Slashes Bug Disclosure Deadlines, Splitting Experts (opens in a new tab)
CyberScoop·
California sues OpenAI over AI agents’ summer hacking spree (opens in a new tab)
The Register - Security·
Infoblox customer and staff data exposed in Sharepoint breach (opens in a new tab)
The Record·
California sues OpenAI for Hugging Face and other incidents, alleging deception and recklessness (opens in a new tab)
Security Affairs·
U.S. Banks Struggle With Outages Amid Wider Technology Disruptions (opens in a new tab)
Bleeping Computer·
Google Project Zero cuts bug disclosures to 30 days amid AI threats (opens in a new tab)
CyberScoop·
Commerce moves to place new restrictions on sale of AI chips abroad (opens in a new tab)
Bleeping Computer·
California sues OpenAI over AI agents escaping sandboxes, hacking firms (opens in a new tab)
The Record·
Match fined $10 million for alleged Tinder privacy violations (opens in a new tab)
SecurityWeek·
Trump Administration Moves to Block China From Indirect Access to American AI Technology (opens in a new tab)
Bleeping Computer·
Barracuda links cloud breach to ARTEX AI-powered cyberattack (opens in a new tab)
Dark Reading·
Ransomware Attacks Increase Behind AI, But Payments Decline (opens in a new tab)
The Register - Security·
Attackers are using AI agents to automate cybercrime at 'industrial scale,' FBI and AWS warn (opens in a new tab)
Dark Reading·
'Sleeper Agent' Tactic Hides Payloads in Open-Weight AI Models (opens in a new tab)
The Register - Security·
‘123456’ has been the world’s most popular password for 7 out of the last 8 years (opens in a new tab)
The Register - Security·
NoisyTCP bug gives Linux servers a remote crash problem 11 years in the making (opens in a new tab)
HighRapid7 Cybersecurity Blog·
Rapid7 Detected Exploitation of CVE-2026-90436 against Atlassian Confluence Data Center, N-able N-central also Exploited (opens in a new tab)
The Record·
Italian regulator fines Meta $886 million for violating data privacy law (opens in a new tab)
The Register - Security·
Ubuntu gets AI-generated security fixes for those feeling brave (opens in a new tab)
The Record·
King Charles opens new HQ for National Cyber Force, UK’s offensive hacking unit (opens in a new tab)
The Register - Security·
Crook turns free trials into phishing factory, as LearnWorlds sends 13,000 fake YouTube copyright notices (opens in a new tab)
Official advisories
- Mitel Releases Security Update for MiCollab (opens in a new tab)
NHS Cyber Alerts· · Alert reference CC-4894
- CISA Adds Two Known Exploited Vulnerabilities to Catalog (opens in a new tab)
All CISA Advisories·
- ConnectWise Releases Security Advisory Following Unauthorised Access to ScreenConnect Environments (opens in a new tab)
NHS Cyber Alerts· · Alert reference CC-4893
- Critical Remote Code Execution Vulnerability in Apache HTTP Server (opens in a new tab)
NHS Cyber Alerts· · Alert reference CC-4892
- Siemens Desigo, Gridscale (opens in a new tab)
All CISA Advisories· · Alert reference ICSA-26-316-02
- Siemens RUGGEDCOM APE1808 (opens in a new tab)
All CISA Advisories· · Alert reference ICSA-26-316-05
Analysis and commentary
- Srsly Risky Biz: Iranian hacks with Russian characteristics (opens in a new tab)
Risky Business News·
- Who Runs the Ransomware Group Qilin? (opens in a new tab)
Krebs on Security·
- The Bitter Wall (opens in a new tab)
Unsupervised Learning·
- New OpenSSH Vulnerability (opens in a new tab)
Schneier on Security·