LLM Prompt Injection Security

8 repos

Defenses, detection systems, and guardrails for protecting large language models against prompt injection attacks and adversarial prompts. This cluster covers security tooling and libraries—primarily in TypeScript and Python—designed to sanitize inputs, validate prompts, and prevent malicious manipulation of LLM behavior. Central projects like Rebuff and HAI Guardrails provide practical frameworks and shields for hardening LLM applications against prompt-based exploits.

Related clusters