A critical vulnerability chain in LangGraph, an open-source framework for building complex AI applications, has been disclosed, allowing for remote code execution on self-hosted AI agents. The flaws, which have now been patched, include a critical SQL injection vulnerability in one of LangGraph's functions1. This vulnerability chain could be exploited by attackers to gain unauthorized access to sensitive data and systems. LangGraph, developed by LangChain, is used to build stateful and multi-agent AI applications, making it a significant target for malicious actors. The patched vulnerabilities highlight the importance of ensuring the security of AI frameworks and applications. The disclosure of these flaws serves as a reminder to practitioners to prioritize the security of their AI systems, as vulnerabilities in these systems can have significant consequences, so what matters most is that developers and users of LangGraph take immediate action to apply the patches and prevent potential attacks.