Sentinely — Security layer for AI agents. Stop prompt injection, memory poisoning, and agent drift in 3 lines of code.
sentinely 0.2.7
Sentinely 0.2.7, a security layer for AI agents, introduced a vulnerability exposing systems to prompt injection, memory poisoning, and agent drift attacks. The flaw allows remote attackers to manipulate AI agent behavior, potentially leading to data exfiltration or unauthorized access, affecting users of the library in production environments.