ml-llm-rag-engineering
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill consists of documentation and architectural guidelines for building secure machine learning systems. It does not include scripts, binaries, or tool definitions that could perform actions on a host system.
- [INDIRECT_PROMPT_INJECTION]: The documentation specifically addresses the risk of indirect prompt injection in RAG (Retrieval-Augmented Generation) systems. It instructs the agent to treat retrieved documents as untrusted input and to use delimiters and explicit instructions to prevent the model from following commands found within retrieved data. While the skill defines the architecture for such systems, it does not provide the tools or scripts that would implement the ingestion surface, and it advocates for robust sanitization and validation measures.
Audit Metadata