darkreading

Researcher shows how agentic AI is vulnerable to hijacking to subvert an agent's goals and how agent interaction can be altered to compromise whole networks.

Initially though to be a DDoS attack, the incident was actually due to a routine change in permissions that caused widespread software failure.

It only takes recycled cans, copper, and cheap gadgets off the Web to trick a train conductor into doing something dangerous.

The collaborative effort combines multiple federal departments, along with private companies to reduce, if not eliminate, billions lost annually to fraud.

IoT devices can be compromised, thanks to gaps in cloud management interfaces for firewalls and routers, even if they're protected by security software or not online.

As vulnerabilities in the Common Vulnerabilities and Exposures ecosystem pile up, one Black Hat Europe presenter hopes for a global, distributed alternative.

Researchers say Israel remains a central focus, with UNC1549 targeting aerospace and defense entities in the US, the UAE, Qatar, Spain, and Saudi Arabia.

Free the logs! Behind the scenes at InfluxData, which turned to its own in-house security monitoring platform, DiSCO, to protect its supply chain after its third-party tool was breached.

A malware campaign presents fake websites that can check if a visitor is a potential victim or a security researcher, and then proceed accordingly to defraud or evade.

Bug bounty programs create formal channels for organizations to leverage external security expertise, offering researchers legal protection and financial incentives for ethical vulnerability disclosure.

Security teams can combat alert fatigue, high rates of false positives, and reduce time spent on manual data collection with Mate's SOC platform, which utilizes AI agents to autonomously investigate and resolve alerts.

The vulnerability could allow an unauthenticated attacker to remotely execute administrative commands.

Four individuals admitted to assisting foreign IT workers in gaining employment at US companies by providing false identities and remote access to employer-owned laptops.

Researchers discovered a security weakness in the AI-powered coding tool that allows malicious MCP server to hijack Cursor's internal browser.

The Akira ransomware group has been experimenting with new tools, bugs, and attack surfaces, with demonstrated success in significant sectors.

A public dataset and platform-agnostic analysis tool aim to help organizations in the fight against Apple-targeted malware, which researchers say has lacked proper attention.

A kitchen-sink approach to building containers has loaded many with vulnerabilities. A handful of companies are trying to slim them down.

A self-replicating attack led to a tidal wave of malicious packages in the NPM registry, targeting tokens for the tea.xyz protocol.

Amazon Web Services VP Sara Duffer highlights the top lessons she brought back to her security role after taking part in Amazon's shadow program.

Most enterprises use more than 1,000 apps, according to ESG research, yet about half are integrated with IGA. Industry innovations enable teams to expand app coverage and get more IGA value.