cerberust
This project's design was inspired by Protect AI's llm-guard
(https://github.com/protectai/llm-guard), a Python toolkit for securing LLM
input and output. cerberust borrows concepts — the scanner/vault model, the
reversible-redaction approach, and the prompt-injection classifier
(protectai/deberta-v3-base-prompt-injection-v2) — but shares no source code with
it. We credit llm-guard as prior art with thanks.