sentinel-prompt-protection

Deploy a state-of-the-art binary classifier using ModernBERT to detect prompt injection attacks and protect LLMs from adversarial input manipulation.

Independent GitHub snapshot · Automated classification · Original documentation governs