knowledge-retrieval
Pass
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [SAFE]: The skill utilizes well-known technology stacks including LlamaIndex, ChromaDB, and NVIDIA Foundational RAG for its intended purpose.- [PROMPT_INJECTION]: Evaluated the indirect prompt injection surface inherent to RAG workflows. Ingestion points: Untrusted data enters the context through ingested documents (PDFs, reports, documentation). Boundary markers: Absent; there are no explicit delimiters to segregate retrieved content from system instructions. Capability inventory: The agent utilizes the
knowledge-retrievaltool. Sanitization: No validation or filtering of external content is specified before the retrieved chunks are provided to the agent.
Audit Metadata