-
A major Claude outage disrupted several Anthropic services on August 16, 2026. Users reported login failures, loading errors and requests that did not complete on Claude.ai, Claude Code and Claude Cowork. The incident began at about 21:58 UTC. Anthropic first identified an authentication problem affecting some users, then confirmed a wider disruption involving degraded performance…
-
Threema DDoS attacks caused major disruptions to the Swiss secure messaging service earlier this week. Users struggled to send messages and connect to the platform as the attacks affected both Threema and its colocation partner. The company said the incidents temporarily made its service unavailable or only partly available on Tuesday evening and Wednesday morning.…
-
A new macOS threat called AmnesiaStealer malware gives attackers an unusual way to abuse stolen browser sessions. Instead of simply stealing cookies and passwords, it can copy a victim’s Chromium browser profile and remotely operate it in a hidden browser instance. The malware spreads through ClickFix attacks that use fake GitHub download pages. Victims receive…
-
Anthropic plans to add invisible watermarking to text generated by future Claude models, making it easier to identify AI-written content. Claude text watermarking will launch worldwide because Anthropic does not yet have a reliable way to limit the feature by region. The move supports the EU AI Act and the EU’s Code of Practice for…
-
Brazilian police have arrested four suspected cybercriminals over an alleged €30 million bank fraud involving unauthorised direct debits from Commerzbank customer accounts. German and Brazilian authorities say the group exploited a vulnerability at a financial service provider. The fraud took place over four days in November 2023 and triggered a major cross-border investigation. Service provider…
-
Hackers are actively exploiting a macOS Screen Sharing flaw to bypass authentication, gain root access and install Monero cryptocurrency miners, according to the Netherlands’ National Cyber Security Centre. Apple fixed the vulnerability, tracked as CVE-2026-65400, on 6 August. However, public exploit code has since emerged, and attackers have started targeting exposed systems. Flaw allows access…
-
Security researchers have detected attacks targeting a maximum-severity SAP Commerce Cloud vulnerability only three days after SAP released a fix. The SAP Commerce Cloud flaw, tracked as CVE-2026-58231, allows an unauthenticated attacker to execute arbitrary code. Defused researchers first spotted exploitation attempts against their honeypots on Friday. Critical flaw enables remote code execution CVE-2026-58231 affects…
-
Shell is investigating a potential cyber incident after the Clop ransomware group claimed it stole 89GB of data from the energy giant. The alleged Shell data theft forms part of a wider campaign targeting internet-exposed PTC Windchill and FlexPLM systems. Clop has listed dozens of organisations on its data leak site and claims to have…
-
The European Data Protection Supervisor has warned that expanded Europol powers could weaken privacy and fundamental rights. The watchdog says lawmakers need to add stronger safeguards before giving the agency broader authority to collect and process personal data. Commission wants to strengthen Europol’s role The European Commission presented its proposed Europol regulation on 24 June…
-
Meta says it has removed 756,000 Facebook and Instagram accounts under Australia’s social media ban for users under 16. The company used AI detection, user reports and account-creation checks, although early data suggests many teenagers continue to access social platforms. Meta removes 756,000 underage accounts Australia’s under-16 social media ban took effect in December 2025.…