# Guided Lab: Exploiting Trust Between Two Cooperating Agents

> A comprehensive, hands-on capstone lab where you will execute a multi-stage attack exploiting inter-agent trust to achieve unauthorized privilege escalation in a simulated multi-agent system.

Canonical: https://codelivly.com/rooms/guided-lab-exploiting-trust-between-two-cooperating-agents
Last updated: 2026-10-14

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

This capstone module provides a fully interactive, guided laboratory environment to synthesize the concepts learned throughout the Multi-Agent & MCP Security learning path. You will perform a multi-stage attack against a simulated enterprise multi-agent system, executing indirect prompt injection, shared memory poisoning, and trust boundary exploitation to achieve unauthorized data exfiltration in a safe, isolated environment.

## Lessons

1. Lab Setup & Two-Agent Target Overview (30 min)
2. Crafting the Cross-Agent Attack Payload (30 min)
3. Verifying the Attack Propagated to the Second Agent (30 min)

## Related

- [Learning path: Multi-Agent and MCP Security](https://codelivly.com/learning-paths/multi-agent-mcp-security)
- [All rooms](https://codelivly.com/rooms)