Exploring vulnerabilities in Azure OpenAI's security measures against prompt injection attacks.