🚨 A LangChain vulnerability let attackers steal OpenAI API keys, prompts & files—just by clicking “Try It.”

All data silently routed through a malicious proxy.

CheckDot is SAFU research on CheckDot .