ContextCrush flaw raises alarms over AI coding tools
Security researchers have disclosed a vulnerability affecting a widely used platform that supplies contextual data to artificial intelligence coding assistants, exposing developers to potential malicious instructions embedded in documentation and external resources. The weakness, labelled “ContextCrush,” was identified by researchers at Noma Labs in the Context7 platform operated by Upstash. Context7 is designed to feed documentation and technical material to AI-powered development tools so they can generate […]