How to meet it
OWASP AISVS v1.01
AI-Assisted Secure Coding (Appendix C) · AI Tool Qualification & Threat Modeling
Threat-model every AI coding tool
Verify that every AI tool, whether it is an assistant, a reviewer, an agent, or an MCP server, has a threat model. The threat model covers misuse, model inversion, training-data leakage, prompt injection from untrusted input, insecure output handling, excessive agency, and risk inherited from its dependency chain.
OWASP AISVS v1.01, requirement AC.2.1
In plain English
Threat-model every AI tool used in development, whether assistant, reviewer, agent or MCP server, before it is adopted.
What good looks like
- Each tool has a threat model covering misuse, data leakage, prompt injection, unsafe output, excessive permissions and risks from its dependencies.
- Threat models are reviewed on major changes.
- Adoption waits for the threat model.
How teams usually do it
- A short threat model template specific to AI tools, filled in during procurement.
Evidence an auditor would accept
- Threat models for tools in use.
Common mistakes
- Threat-modelling the code assistant but not the MCP servers developers add to it.
Not sure which requirements apply to your system? Take the assessment.
Requirement text from the OWASP AI Security Verification Standard, licensed under CC BY-SA 4.0. This guidance is shared under the same licence.