# What is System Prompt Extraction

> Understand the mechanics of system prompt extraction. Learn why attackers target the "DNA" of AI applications and how to defend against prompt leaking.

Canonical: https://codelivly.com/rooms/what-is-system-prompt-extraction
Last updated: 2026-10-09

| | |
|---|---|
| Level | Intermediate |
| Category | Cybersecurity |
| Lessons | 2 |
| Estimated time | 30 min |
| XP | 100 |

This module introduces System Prompt Extraction (OWASP LLM06), a critical reconnaissance technique where attackers force an LLM to reveal its hidden foundational instructions. You will learn the mechanical differences between jailbreaking and prompt leaking, how the LLM's attention mechanism blurs the line between instructions and data, and how defenders can implement the "Sandwich Defense" and Instruction Hierarchy to protect proprietary AI logic and hardcoded secrets.

## Lessons

1. What is a System Prompt and Why It's Considered Sensitive (30 min)
2. Why Extraction is Possible Even When "Hidden" (30 min)

## Related

- [Learning path: System Prompt and Data Extraction](https://codelivly.com/learning-paths/system-prompt-and-data-extraction)
- [All rooms](https://codelivly.com/rooms)