How to meet it
OWASP AISVS v1.01
Input Validation · Content & Policy Screening
Detect attacks that mix input types
Verify that coordinated attacks spanning multiple input types (e.g., steganographic payloads in images combined with prompt injection in text) are detected and blocked.
OWASP AISVS v1.01, requirement 2.2.4
In plain English
Detect attacks that are split across input types, such as part of an instruction in an image and the rest in the text, which look harmless when each part is checked alone.
What good looks like
- Checks look at the combined input, as the model will see it, after extraction from each media type.
- Test cases include attacks split across modalities.
- Detected attempts are blocked and logged.
How teams usually do it
- Extracting text from all inputs and screening the assembled prompt as one.
- Red-team exercises that combine media types.
Evidence an auditor would accept
- Pipeline design showing where the combined check happens.
- Results from multi-modal attack tests.
Common mistakes
- Separate filters per input type with no check of the combination.
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.