# Agent Memory Poisoning and Persistence Attacks

> Master the techniques of agent memory poisoning and persistence attacks, learning how to embed dormant payloads in long-term memory for delayed, stealthy exploitation.

Canonical: https://codelivly.com/rooms/agent-memory-poisoning-and-persistence-attacks
Last updated: 2026-10-13

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

This module explores the critical vulnerabilities in how AI agents store and retrieve state. You will learn how to execute memory poisoning attacks, embedding dormant, malicious instructions into an agent's episodic or semantic memory. Discover how these payloads lie in wait to hijack future, unrelated interactions, and master the defensive strategies required to sanitize, isolate, and audit agentic memory stores.

## Lessons

1. What is Memory Poisoning in an Agent Context (30 min)
2. Planting Malicious Instructions in Long-Term Memory (30 min)
3. Why Poisoned Memory Persists Across Sessions (30 min)

## Related

- [Learning path: Agent Exploitation](https://codelivly.com/learning-paths/agent-exploitation)
- [All rooms](https://codelivly.com/rooms)