realtime-risk-engine
Pass
Audited by Gen Agent Trust Hub on Sep 3, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external CSV files, which creates a potential surface for indirect prompt injection where instructions could be hidden in data fields.
- Ingestion points: The
scripts/realtime_risk_engine_diagnostics.pyscript reads external data usingpd.read_csv(arguments.input_csv)at line 62. - Boundary markers: The instructions lack explicit boundary markers or directions for the agent to treat CSV content strictly as data and ignore any embedded natural language instructions.
- Capability inventory: The diagnostic script has file-system write capabilities, allowed via the
--outputflag which writes a JSON file to a user-specified path (line 72). - Sanitization: The script performs data sanitization for specific numeric columns using
pd.to_numeric(..., errors="coerce")at line 21, ensuring only numeric data is used in statistical calculations, though the initial loading of the file into memory does not filter non-numeric content.
Audit Metadata