rag-systems
Pass
Audited by Gen Agent Trust Hub on Nov 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely instructional and does not contain any executable scripts (Python, JavaScript, or Shell). It serves as a knowledge base to guide the agent in designing RAG pipelines.
- [DATA_EXFILTRATION]: No network operations or data exfiltration patterns were detected. The skill references several external services (Cohere, OpenAI, Qdrant) as part of architectural recommendations, but does not perform any calls to them.
- [CREDENTIALS_UNSAFE]: No hardcoded API keys, tokens, or sensitive credentials were found. It correctly suggests using configuration files or environment variables for system specifications.
- [REMOTE_CODE_EXECUTION]: There are no patterns involving the download or execution of remote code. All technical steps are provided as text-based instructions for system design.
- [INDIRECT_PROMPT_INJECTION]: While the skill defines systems that ingest untrusted data (PDFs, Markdown), it proactively includes a verification checklist recommending that models be instructed to cite sources and refuse to answer if context is absent, which is a key mitigation strategy.
Audit Metadata