Together.ai Prompt Leakage via Exposed Repository Credentials
November 21, 2026
An unauthorized party downloaded prompt samples for MiniMax M3, GLM-5.2, and Kimi K2.6 models after an AI coding agent committed a credential to a public GitHub repository. The exposure occurred via a private Hugging Face repository containing samples collected during June and July 2026.
HOW THIS AFFECTS YOU
●
builderYou need to audit your AI agents' access to secrets to prevent automated credential leaks.
●
policyThis shows how agentic workflows can create new data privacy and security risks.