How to use this reference
For an owned RAG design, document dataset partitions and permission-aware retrieval, preserve source classifications when combining knowledge, and review ingestion integrity. Record retrieval events so confidentiality and provenance decisions remain reviewable.
Before reading
- Basic concepts of embeddings, vector stores and retrieval-augmented generation
- User, group and tenant access-control models
Context and limits
- The 2025 designation is an edition identifier. The edition-2025-pdf source dates its release to November 18, 2024, while edition-2025-landing displays November 17, 2024 for that landing page. Neither date establishes the individual web entry publication or latest update; those remain unknown.
- As of October 6, 2026, release-index labels the 2025 edition archived and the 2026 edition current; edition-2026-status confirms the newer release is published. This record preserves its 2025 URL, identifier and educational summary. Technical equivalence with the 2026 edition has not been assessed.
- This record focuses on permissions and provenance; it does not reproduce the source’s attack scenarios or quantify embedding reconstruction risk.
- RAG grounding can improve relevance without establishing that retrieved content is authorized or trustworthy.
Sources
- LLM08:2025 Vector and Embedding WeaknessesOWASP Gen AI Security Project · reviewed 2026-10-03
- OWASP Top 10 for LLM Applications 2025: official edition PDFOWASP Gen AI Security Project · reviewed 2026-10-06
- OWASP Top 10 for LLM Applications 2025: resource landing pageOWASP Gen AI Security Project · reviewed 2026-10-06
- OWASP LLM Top 10 legacy repository: current and archived releasesOWASP · reviewed 2026-10-06
- OWASP LLM Top 10 2026: published release overviewOWASP GenAI Security Project · reviewed 2026-10-06
Record reviewed 2026-10-06. Snapshot a066321c166a. Open the complete JSON contract.