SECAI Core · phase 11 of 15
Spoofing, tampering, repudiation, information disclosure, DoS, and elevation of privilege
Threat model developed by Microsoft.
The Explain card
- Plain English
- STRIDE is Microsoft's threat modelling framework. It names six threat categories: spoofing, tampering, repudiation, information disclosure, denial of service, and elevation of privilege, and walks each component through them.
- Example
- Threat modelling an LLM agent with STRIDE, a team asks: can a tool output spoof a user message? Can a document tamper with instructions? Can the model leak the system prompt? Can a prompt escalate its permissions?
- Why it matters
- STRIDE gives AI defenders a structured checklist so they stop reviewing only the headline risks and cover the quieter failure modes too.
- Hook
- Six doors on every component; walk through each one.
Where it sits in the deck
Phase 11: Threat Landscape: Attack Vectors and Adversarial Techniques
With defences named, learn what they must defend against — the full catalogue of attack techniques targeting AI systems, their inputs, outputs, training pipelines, and supply chains.
