AI slopsquatting is the practice of registering package names that code assistants or other AI tools may hallucinate or suggest, then using those names as malicious dependencies. The goal is to exploit trust in a believable library name before the developer checks whether it is real or safe.
This matters because modern software often installs packages directly from public registries such as npm. If an AI recommends a nonexistent, misspelled, or confusingly similar package, an attacker can publish that name first and turn a simple install into a supply-chain compromise. Defenders should verify that suggested packages actually exist, compare the name against known libraries, review maintainers and provenance, and use allowlists or approval gates for new dependencies. In practice, slopsquatting attacks succeed at the moment a dependency is trusted, often before any malicious code is executed.