prompt-protection

Detect and strip malicious prompt injection, jailbreaking, and data exfiltration attempts from LLM inputs. Zero dependencies. Works in Node.js and browsers.