Back to today’s briefing

Story perspectives

Cybersecurity Alert: LLMs Fabricate Dangerous Code Packages

4/15/2025

25 5

Summary

Story summary

Researchers have raised alarms about large language models (LLMs) that can fabricate non-existent code packages, a dangerous tactic known as "slopsquatting." In tests, almost 20% of suggested packages were illusions. This opens the door for cybercriminals to create harmful duplicates, jeopardizing software integrity. Developers must now double-check package authenticity to safeguard their work.

1 of 1

Story summary

Researchers have raised alarms about large language models (LLMs) that can fabricate non-existent code packages, a dangerous tactic known as "slopsquatting." In tests, almost 20% of suggested packages were illusions. This opens the door for cybercriminals to create harmful duplicates, jeopardizing software integrity. Developers must now double-check package authenticity to safeguard their work.