am-literature-rag
Pass
Audited by Gen Agent Trust Hub on Nov 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The
scripts/amrag.pyscript uses thesubprocessmodule to call thepdftotextbinary (part of the Poppler suite) for extracting text from PDF files. The implementation follows security best practices by passing arguments as a list rather than a shell string, preventing shell injection attacks. - [EXTERNAL_DOWNLOADS]: When using the 'dense' or 'hybrid' backends, the skill downloads the
all-mpnet-base-v2embedding model from Hugging Face using thesentence-transformerslibrary. This is a standard operation for local embedding generation and uses a well-known service. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external data (user-provided PDFs). While this introduces a surface for indirect prompt injection—where malicious instructions hidden in a paper could attempt to influence the agent—the skill includes specific instructions to the agent to ground all answers strictly in the retrieved citations and to explicitly label general knowledge.
Audit Metadata