AI reads 36 sources every 2 hours and ranks each story by importance from 1 to 5 - so you see what actually matters.
The Hacker News examines how the rapid push to deploy AI systems has far outrun the security practices needed to protect them, leaving organizations exposed. The piece argues that security work is trailing AI development by years or decades, and that closing the gap requires treating safety as a core part of AI engineering rather than an afterthought.
A vulnerability in GoBalance, a Go library for ed25519 key operations, allows attackers to recover Tor-format private keys from generated addresses. With a stolen key, an adversary can impersonate or hijack a .onion service, potentially intercepting traffic and impersonating hidden sites. Tor operators relying on the affected library should rotate keys and apply available fixes.
At the Pwn2Own hacking contest, three separate teams each demonstrated remote exploits against Google Pixel 10 phones running the latest, fully patched software. The attacks succeeded without physical access or user interaction, exposing previously unknown flaws in one of Google's most secure devices. Such findings are typically disclosed privately to the vendor for fixing, but they highlight ongoing risks to Android users even on up-to-date hardware.
The FBI has seized seven domains tied to Flax Typhoon, a China-linked hacking group, disrupting infrastructure the group used to break into critical networks. The takedown targets tooling used in intrusions against US critical infrastructure, marking a significant blow to a state-sponsored threat actor.
Anthropic has launched a free security scanning service that uses AI to find vulnerabilities in open-source software. The move targets widely used open-source code, which underpins much of the internet but often lacks dedicated security funding. It could help catch flaws before attackers exploit them, though it also expands Anthropic's role in developer tooling.
The FBI says hackers tied to China operated an online portal that let outside parties browse and download emails stolen in earlier intrusions. The disclosure points to a formalized, brokered distribution of stolen data rather than isolated espionage. It underscores the scale of state-linked cyber-espionage and the risks to governments and private organizations whose breached correspondence may still be circulating.
The Hacker News roundup reports a ransomware affiliate betrayal, a WhatsApp remote access trojan, exposed hacker tools, and 12 additional cybersecurity stories. It highlights ongoing threats involving ransomware crews, mobile malware, and leaked attack tools. The digest matters because it gives defenders and users a quick view of active security risks.
Japan has seen a sharp increase in web data leaks, driven largely by abuse of mobile application APIs and attacks targeting Metabase, the open-source business intelligence tool. Attackers are exploiting poorly secured API endpoints and misconfigured Metabase instances to reach sensitive data at scale. The trend matters because API and analytics-tool weaknesses are common worldwide, so the same attack patterns could affect organizations well beyond Japan.
A cyberespionage campaign by UAC-0099 is targeting Ukrainian government personnel with the ASHVEIN remote access trojan, which hides malicious commands in HTML. The malware can give attackers persistent access, enabling data theft and espionage against government networks. It matters because it highlights continued cyber threats against Ukraine and evolving evasion techniques.