rag-system
Pass
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill implements a RAG system that is susceptible to indirect prompt injection through its data processing logic. \n
- Ingestion points: Untrusted external document content and user queries are ingested into the system context in the
optimize_contextandretrievefunctions withinSKILL.md. \n - Boundary markers: The current implementation lacks clear delimiters or explicit instructions to the agent to disregard potential commands embedded within retrieved content. \n
- Capability inventory: The skill manages vector storage, embedding generation, and context retrieval. \n
- Sanitization: There is no evidence of sanitization, filtering, or validation of the retrieved content before it is concatenated into the agent's context window.
Audit Metadata