Software developers' use of large language models (LLMs) presents a bigger opportunity than previously thought for attackers to distribute malicious packages to development environments, according to ...