CNCERT warns OpenClaw AI agent has weak defaults enabling prompt injection and data leaks, prompting China to restrict use on government systems.

Qualys researchers expose ‘CrackArmor’ flaws that allow unprivileged users to escalate privileges to root, break container isolation, and crash systems, with no CVE identifiers yet assigned.

Two critical-severity n8n vulnerabilities could have led to unauthenticated remote code execution, sandbox escape, and credential theft.

Direct prompt injection occurs when a user crafts input specifically designed to alter the LLM’s behavior beyond its intended boundaries.

AWS launched managed OpenClaw on Lightsail for AI agent deployment while security concerns mount. The 250k-star GitHub project is affected by CVE-2026-25253, which enables one-click RCE, with 17,500+ ...