Cybersecurity

Bro researchers dropped a strat warning about AI agents. one agent could potentially influence another by passing malicious instructions through shared files. that's the risk we're looking at.

AI agent security risk alert

Content creators should keep their account-opening email separate from the one they list publicly. I’ve spotted phishing emails just by noticing they landed in my contact inbox instead of the original sign-up address,… Mostra altro

Email security tip for creators

Anthropic's Claude Mythos AI found novel methods to break cryptographic algorithms, including a post-quantum encryption candidate

Anthropic Claude Mythos AI cryptographic attack

UK AI Safety Institute found deceptive behavior in several advanced AI models during cybersecurity tests

AI deceptive behavior UK safety report

Anthropic hanging out with some cybersecurity analysts rn

The idea of an ai breaking out of its sandbox and hacking into secure websites is honestly kinda terrifying. makes you wonder what happens when they get really smart and start going after things like nuclear codes or… Mostra altro

The fbi flagged five games on steam for having malware that stole over $220,000 from users. thousands of gamers were impacted by these infected titles

fbi warns of steam game malware

The FBI has arrested a hacker who hid cryptocurrency-stealing malware inside Steam video games. The malware was designed to steal passwords, personal data, and drain victims' crypto wallets after installation.

FBI arrests hacker for crypto malware in Steam games

An open-source tool that transforms a local AI into an autonomous hacker has reached 4,000 stars on GitHub. T3MP3ST, created by elder_plinius, is gaining significant attention. Described as a multi-agent… Mostra altro

T3MP3ST GitHub 4,000 stars

The most common passwords of 2024 have been revealed, and the list is as predictable as ever. According to an infographic from NordPass, the top 25 includes classics like "123456" at number one, followed by "123456789"… Mostra altro

World's Top 25 Most Common Passwords 2024

You can change your IP address as much as you want, but the underlying Windows installation still carries a unique identifier. That's the subtle takeaway from the Peter Stokes case. Stokes, a 19-year-old dual… Mostra altro

Peter Stokes Scattered Spider case

A 15-year-old in Japan was arrested for allegedly using ChatGPT to assist in developing a program that was used in a cyberattack on Bandai Namco Filmworks. According to police, the attack caused 46,812 users to be… Mostra altro

Bandai Namco cyberattack teen arrested

A user's two-factor authentication screen accidentally displayed the verification code in plain text, defeating the purpose of the security measure.

2FA code leaked in plain text

Anthropic has released a 36-page security guide that essentially advises enterprises to stop trusting their own AI agents. The document, which covers topics like zero trust for AI agents, is especially relevant for… Mostra altro

Anthropic zero trust AI agents guide

The U.S. government has just given the green light for Claude Mythos 5 to be redeployed. This is the cybersecurity model from Anthropic that was banned on June 12,the most powerful one they have ever created. The NSA… Mostra altro