OpenClaw AI Agents Weaponized to Auto‑Distribute Malware via Malicious Skills

VirusTotal researchers have uncovered that threat actors are hijacking OpenClaw AI agents and publishing malicious “skills.” These skills instruct compromised endpoints to fetch and execute untrusted code, turning the AI platform into an automated infection vector that requires no direct human interaction.

The abuse creates a highly scalable, stealthy supply‑chain style attack that can bypass many traditional defenses because the malicious payload is delivered through a legitimate AI service. Defenders must start monitoring OpenClaw traffic, enforce strict validation of downloaded code, block unknown skill execution, and integrate these indicators into detection and response playbooks.

Categories: AI Security & Threats, Malware & Ransomware, Threat Intelligence, #AI Security & Threats

Source: Read original article