How to meet it
OWASP AISVS v1.01
Memory, Embeddings & Vector Database Security · Embedding Sanitization & Validation
Reject content built to game search
Verify that content crafted to manipulate retrieval results is detected and rejected or quarantined before vectorization.
OWASP AISVS v1.01, requirement 8.2.4
In plain English
Detect content written to manipulate search results, such as keyword-stuffed pages or hidden instructions, before it is indexed.
What good looks like
- Ingested content is checked for manipulation patterns.
- Suspicious content is rejected or quarantined.
- External sources get stricter checks.
How teams usually do it
- Classifiers for injection and spam at ingestion.
- Limits on how many chunks from one source can be retrieved together.
Evidence an auditor would accept
- Ingestion checks and test results.
Common mistakes
- Indexing public web content with no checks.
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.