Slopsquatting: AI coding tools spawn new supply chain attack vector

Slopsquatting represents an emerging class of supply chain threat enabled by the widespread adoption of AI coding assistants. Unlike traditional typosquatting, which depends on human typographical errors or inattention, slopsquatting exploits a fundamental characteristic of large language models: their tendency to generate plausible-sounding but entirely fictitious package names. The attack is particularly insidious because developers trust the code suggestions generated by their AI tools, making them unlikely to scrutinize whether a package name is real.

The threat model is straightforward but effective. When an LLM suggests a nonexistent open-source package during a coding session, developers may install it without verification. If a threat actor has already registered that exact fictitious package name and seeded it with malicious code, the developer's system becomes compromised from the outset. This grants attackers early and deep access to the software supply chain, potentially affecting not only the initial developer but also downstream users of any code or product built on the compromised foundation.