Agent Security
AI Agent Velocity and Supply Chain Gaps Define Current Security Risks
Rapid code merges and unverified hardware imports highlight the need for stricter oversight in automated AI deployments and procurement channels.
Cyber Attacks
View allCyber Attacks
GhostAction Campaign Infects 30,000 GitHub Repos to Steal Cloud Keys
A renewed supply chain attack pushes malicious workflows into thousands of repositories, harvesting cloud and AI credentials from source code and git history.
Cyber Attacks
Vercel rolls out device-bound sessions to neutralize stolen cookies
Vercel introduces Device Bound Session Credentials to limit infostealer impact, launching alongside an AI security review tool for Pro and Enterprise teams.
Cloud Security
View all01
Spotting Exposed Storage Buckets: A Practical Training Plan
02Business: Teams should note that Microsoft is testing phone-as-controller and Wi-Fi Nano support, features
03Cyber Security News Ranks Top 10 Identity Governance Tools for 2026
04Cloud supply chain risks on personal devices and how to stop them
05484,142 Exposed Cloud Buckets Hold 18 Billion Files, Research Finds
06Wiz CIRT sees AI agents driving hundreds of cloud actions per hour
Agent Security
View allAgent Security
Adobe and Disney Partner to Build Custom AI Models on Characters
The multi-year deal expands training data to include Disney characters, settings and styles for private model development.
Agent Security
Two Approved Nvidia H200 Buyers Found With Empty Offices and Sanctioned Links
Investigation reveals U.S.-cleared chip importers have vacant premises and ties to firms restricted by Washington for military technology transfers.
Agent Security
AI infrastructure commitments outpace verified economic value
Contracted demand does not guarantee productive utilization, creating verification gaps in long-term capacity agreements and data-center builds.
Latest news
Vulnerabilities
CVE-2026-111792: Update Apache Batik to 1.20 to Fix DoS Flaw in SVG Element Reference Resolution
A new update resolves a denial of service vulnerability in Apache XML Graphics Batik that affects version 1.19.
Vulnerabilities
CVE-2026-47591 Allows Unprivileged Users to Gain Root in GPU Clouds
Researchers disclosed a local privilege escalation flaw in multi-tenant GPU environments that bypasses standard kernel hardening mechanisms.
Vulnerabilities
Node.js Patches 16 Flaws Including HTTP/2 Crashes and Memory Leaks
Node.js releases updates for 22.x, 24.x, and 26.x lines to fix 16 vulnerabilities, including high-severity issues that crash servers or leak memory.
Vulnerabilities
CVE-2026-126176: Unauthenticated SAML Injection in Fortinet Log and Policy Managers
A high-severity XML injection flaw in FortiAnalyzer and FortiManager allows attackers to alter signed SAML responses without credentials.
Vulnerabilities
CVSS 8.6: iSulad Symlink Bug Lets Containers Steal Host Data via Malicious Images
A symlink vulnerability in openEuler iSulad allows containers to read sensitive host data via crafted images, affecting versions up to 2.1.6.
Vulnerabilities
CVSS 8.6: Webmin 2.670 Flaw Lets Delegated Admins Execute Root Commands via Module Manipulation
A flaw in Webmin versions up to 2.670 allows delegated administrators to grant themselves root-level command execution via module manipulation.
Vulnerabilities
CVSS 8.8: Dell Terraform Providers for ObjectScale and PowerStore Leak Secrets in Logs
Dell Terraform providers for ObjectScale and PowerStore leak sensitive info in logs, rated CVSS 8.8 HIGH, requiring immediate version checks.
Vulnerabilities
CVE-2026-0325: Authenticated attackers bypass IP filters in Palo Alto GlobalProtect
Palo Alto Networks advises that authenticated attackers with valid credentials can circumvent source IP filters via CVE-2026-0325.