Together AI discloses prompt-sample leakage after credential exposure
Together AI said an unauthorized party used an exposed repository credential to download stored prompt samples from its serverless inference service. The samples were collected for authorized software testing and came from requests to the MiniMax M3, GLM-5.2, and Kimi K2.6 models, primarily submitted on June 24 and July 4, 8, and 14, 2026. Together AI said the unauthorized party searched the samples for other service credentials; five tested credentials were canaries associated with a Dragos security experiment. The company has not identified evidence of additional prompt leakage, but its review of the affected contents remains ongoing to determine whether sensitive personal information or other customer data was exposed and which customers need direct notification.
The incident shows how an AI coding agent can commit a credential that exposes authorized testing prompts and potentially sensitive customer data. The scope of that exposure remains under review.
We are notifying customers about a security incident involving a subset of prompts submitted to Together AI’s serverless inference service in June and July 2026. An unauthorized party downloaded stored prompt samples after a credential for a private repository was exposed in a public code repository.
We apologize for this incident. We are reviewing the affected data to determine what information was exposed and which customers require direct notification.
What happened
As part of routine software testing, we collected prompt samples authorized for that use and stored them in a private Hugging Face repository. The samples came from requests to the MiniMax M3, GLM-5.2, and Kimi K2.6 models, primarily on June 24, July 4, July 8, and July 14, 2026.
During development of our public GitHub repository, xorl, an AI coding agent committed a credential that provided access to the private Hugging Face repository. Our automated scanning detected the exposed credential, and we reverted the commit.
During the exposure window, an unauthorized party obtained the credential and used it to download the stored prompt samples. The party subsequently searched those samples for credentials to other services and tested whether those credentials were valid. Five of the credentials tested were canary credentials associated with a Dragos security experiment.
#Together AI#MiniMax M3#GLM-5.2#Kimi K2.6#Hugging Face#xorl#Dragos