ai-security
Sandboxing an AI Agent Is Not Enough: Monitor, Detect, Contain
In July 2026 an AI evaluation harness broke containment and breached a real company. Prevention alone will not hold an agentic model. Here …
Cybersecurity Research & Technical Articles
Breaking things to understand them better
# Latest Articles
View all →
linux-security
Linux LPE PoCs: when the page cache becomes the exploit
A human-readable walkthrough of the linux-lpe-pocs repository, the recent wave of Linux local privilege escalation bugs, why page-cache …
security-operations
Vulnerability Management: Where to Start and Why Automation Is Non-Negotiable
With 45,777 CVEs published in 2025 and NIST NVD falling behind on enrichment, manual vulnerability management is a losing battle. Learn how …
linux-security
Ubuntu Hardening: Automation is one of the key
A default Ubuntu installation is not production-ready. Learn how to harden it systematically using CIS Benchmarks, Ansible automation, …
linux-security
SSH Hardening in 2026: From fail2ban to Certificate-Based Authentication
Most SSH setups are stuck in 2015. Learn how to move beyond password authentication and static keys to a modern hardened configuration with …
application-security
React2Shell (CVE-2025-55182): is there any prevention mechanism?
CVE-2025-55182 gave attackers pre-auth RCE through React Server Components with a single HTTP POST. Learn the root cause, related CVEs, WAF …