LAB 378 – Defending Node.js AI Applications from System Prompt Leakage (NEW)

Course Overview

System Prompt Leakage exposes hidden LLM instructions to users, allowing them to predict, manipulate, or subvert model behavior and bypass security controls. Aligned with OWASP Top 10 LLM 2025 this lab provides hands on experience identifying and mitigating System Prompt Leakage in Node.js AI applications by isolating prompts from user access, obfuscating or randomizing instructions, and blocking unauthorized access to internal prompt structures.

By the end of this skill lab, learners will be able to:

• Identify and test for System Prompt Leakage vulnerabilities
• Apply isolation and obfuscation techniques to system prompts
• Prevent unauthorized access to internal instructions