Resources
AI Agent Threat Scenarios & Control Questions
Use generalised scenarios to identify what your own deployment needs to prevent.
Scenarios
MCP tool poisoning
A malicious or compromised tool can influence the agent before execution.
Prompt injection and private data leakage
Untrusted content can attempt to redirect an agent’s retrieval and export behaviour.
AI-assisted cyber operations
Automation can connect reconnaissance, tool use and repeated attempts.
Start with a clear scope
Tell us which systems, actions and review requirements are in scope. We will discuss the work, responsibilities and deliverables before you commit.