Crypto Guards LLM Prompts and Context
Proposes authenticated prompts and context for cryptographic provenance in LLM apps. Features policy algebra with Byzantine resistance and layered defenses. Achieves 100% attack detection with zero false positives.






