scenario · 11 MIN
Treat model output as untrusted input
Build features on a model without letting its output, or text it reads, take control of your system.
HOOK
The model is not your code. Treat what comes out of it the way you treat what comes from a stranger.
Checking your account…
SOURCES
- LLM01:2025 Prompt Injection — OWASP Gen AI Security Project
- LLM05:2025 Improper Output Handling — OWASP Gen AI Security Project
- Secrets Management Cheat Sheet — OWASP Cheat Sheet Series