LLM Protector
LLM vulnerability scanner
Security scanner that probes a local LLM for prompt-injection and jailbreak vulnerabilities — 50 attacks across 11 categories, grading each model with a severity-weighted risk score. A headless benchmark pipeline compares multiple models; the live demo replays a real recorded scan instantly. Built with AI-assisted development.
Category
Languages
Topics
GitHub
Started Last updated
June 2026