Capture Notes

Paper proposing a layered security framework against prompt injection in RAG-based chatbots.

AI security relevance:

• Useful for RAG guardrail architecture comparison.
• Should be assessed against newer detector calibration and RAG poisoning pipeline papers.