Scan any LLM chatbot for vulnerabilities. Built by Mozilla (github.com)

• Framework for identifying LLM attack vectors
• Targeted testing for prompt injection and data privacy
• Standardizing AI security protocols

"Mozilla has introduced a new open-source framework designed to scan Large Language Model (LLM) chatbots for security vulnerabilities. The tool enables developers to test for common threats such as prompt injection, data leakage, and harmful output generation, aiming to bolster the resilience of generative AI deployments."