1 story · sorted newest first · 📡 RSS
Recent incidents expose decades-old shell injection risks in AI coding agents, with MCP identified as a key vulnerability. The Pen