Malicious TanStack and Guardrails-AI Packages Trigger Supply-Chain Credential Theft

Attackers abused trusted open-source distribution channels to publish malicious package updates, including 84 tainted TanStack releases uploaded within six minutes and a malicious payload discovered in the Guardrails-AI package on PyPI. The compromised packages were built to steal credentials from infected developer devices and move laterally to connected systems, highlighting how software supply-chain attacks can rapidly spread through widely used development dependencies.

OpenAI confirmed that employee devices were affected by the TanStack incident, giving attackers limited access to a small number of internal code repositories. The company said there was no evidence that customer data, production systems, intellectual property, or software releases were compromised, and it reported no persistent threat in operational infrastructure. Researchers detected the malicious activity in about 20 minutes, likely constraining the blast radius, while OpenAI rotated cryptographic certificates associated with the affected repositories as part of containment.