Terabytes of credentials leaked in massive supply-chain attack (arstechnica.com)

A massive supply-chain attack targeting LiteLLM, an open-source AI development tool, has exposed terabytes of sensitive credentials belonging to high-profile organizations like Microsoft, Amazon, Cisco, and Salesforce. Security firms CloudSEK and Hudson Rock revealed that the breach occurred during a 40-minute window in March, where users inadvertently downloaded compromised versions of LiteLLM from the official Python Package Index repository. The leaked data includes critical access secrets such as cloud keys, repository tokens, SSH keys, and Kubernetes secrets, potentially granting attackers access to over 2,500 organizations. This incident underscores significant vulnerabilities in software supply chains, particularly in the fast-evolving AI landscape, where rushed development often compromises security. The breach is linked to a prior attack on the widely used vulnerability scanner Trivy, which has implications for software integrity across numerous tools. The group responsible, TeamPCP, predominantly comprised of teenagers, highlights the growing threat posed by non-traditional attack vectors, emphasizing the need for stronger security measures and rigorous DevOps practices to protect sensitive information in the AI and tech sectors.