# Multi-Turn and Context-Building Injection Attacks

> Learn how attackers use multi-turn conversations to gradually bypass AI guardrails through context manipulation and semantic grooming.

Canonical: https://codelivly.com/rooms/multi-turn-and-context-building-injection-attacks
Last updated: 2026-10-08

| | |
|---|---|
| Level | Intermediate |
| Category | Cybersecurity |
| Lessons | 3 |
| Estimated time | 30 min |
| XP | 100 |

This module explores advanced prompt injection techniques that unfold over multiple conversational turns. You will learn how attackers use context-building, semantic drift, and persona reinforcement to slowly erode an LLM's safety guardrails, bypassing single-turn intent classifiers and setting up devastating payloads.

## Lessons

1. What is a Multi-Turn Attack and Why Single-Message Filters Miss It (30 min)
2. Building Malicious Context Across a Conversation (30 min)
3. Combining Direct and Indirect Techniques (30 min)

## Related

- [Learning path: Prompt Injection Fundamentals](https://codelivly.com/learning-paths/prompt-injection-fundamentals)
- [All rooms](https://codelivly.com/rooms)