How to Prevent Prompt Injection: Why Pre-LLM Sanitization Matters

Prompt injection is a security vulnerability where untrusted input is interpreted as instructions by an LLM, allowing attackers to override system behavior. Effective prevention requires pre-LLM sanitization of user input through validation and filtering, ideally using code-level static analysis rather than just runtime filters.