The now patched vulnerabilities in the rapidly growing AI agent framework allow attackers to steal credentials, escalate privileges, and maintain persistence.
The release of Shai-Hulud source code spells trouble for software developers as researchers worry the self-replicating worm could scale.
Security experts have long warned that insecure automatic tank gauge (ATG) systems exposed on the Internet can be tampered with by threat actors.
AI agents capable of discovering and exploiting obscure vulnerabilities are emerging alongside developers producing vast amounts of potentially flawed AI-generated code, forcing defenders to adapt accordingly.
From CrowdStrike's outage and SIEM's longevity to epic business fails and the post-breach jaded reality, Dark Reading looks back at the mistakes, miscalculations, systemic failures, and cringeworthy moments that still have us shaking our heads.
South Korea's local elections next month will be a test bed for how effective regulations might be to stymie the flow of deepfakes.
Robert "RSnake" Hansen, Katie Moussouris, Rich Mogull, Richard Stiennon, and Bruce Schneier reflect on how their favorite columns penned for Dark Reading over the past 20 years have stood the test of time.
A Taiwanese student experimenting with software-defined radio technology shut down three bullet trains for nearly an hour, leading to an anti-terrorism response.
The new acquisition looks to boost visibility into third-party ecosystems that are becoming a bigger concern as vectors for supply-chain attacks.
This is the second time this year a threat actor has leveraged a CVSS 10.0 vulnerability in Cisco's network control system.
The House Committee on Homeland Security sent a letter about the Canvas cyberattack, the same day that the edtech company said it reached an "agreement" with the ShinyHunters cybercriminals.
In a role reversal, investment dollars in security startups exceeded the value of mergers and acquisitions in 1Q26 by more than $1 billion, a rare occurrence.
A Nitrogen ransomware attack on Foxconn's North American facilities is one of 600 hits on manufacturers this year, as gangs increasingly target the sector for its low tolerance for downtime.
Security governance needs to be more than an annual compliance exercise. New companies are emerging to address risk-management gaps in current audit tools.
Threat actors are publishing RubyGems packages that include scrapers targeting public-facing UK government servers, but with no clear objective.
An OPSEC failure provides a window into what helped the ransomware group rise: a generous affiliate model, opportunistic TTPs, and an effective organizational structure.
Informa TechTarget's flagship cybersecurity media brand launches a special content series to mark two decades as a trusted source for cybersecurity professionals.
Attackers uniquely fingerprint victims before delivering spear-phishing payloads aimed at espionage, in the latest campaign from the Belarussian nation-state threat group.
The cyberthreat group targets an Azerbaijani oil and gas firm with repeated attacks, as the China-linked actors extend targeting beyond hospitality, telecom, and government sectors.
In the latest evolution of automated cyberattacks, threat actors heavily leveraged AI agents to support campaigns against entities in Mexico and Brazil.
It's the first time in two years with no zero-days. But with 137 flaws to patch, including nine critical ones, admins still have plenty of work to do.
A tokenizer library file present in Hugging Face AI models can be manipulated to hijack the model's outputs and exfiltrate data.
As part of Dark Reading's 20th anniversary special coverage, we profile the CISOs, founders, researchers, criminals, and policymakers who rewrote the enterprise risk playbook.
Hundreds of npm packages infected by the self-propagating, credential-stealing worm from TeamPCP are related to the open source TanStack ecosystem.
The Federal Communications Commission eased some restrictions and pushed back deadlines for foreign router manufacturers, but the ban is still in place.
Security controls can do only so much. Here are four attacks where your employees are usually your first, and only, line of cyber defense.
The privilege escalation vulnerability, which is similar to other Linux flaws like Copy Fail and Dirty Pipe, may already be under limited exploitation.
Cyber adversaries have long used AI, but now attackers are using large language models to develop exploits and orchestrate complex attacks.
The campaign quietly compromises aerospace and drone operators to exfiltrate GIS files, terrain models, and GPS data and gain a clear picture of adversaries' world view.
The edtech company is struggling to wrest control from its hackers. PII belonging to hundreds of millions of people is on the line.