day-trading
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from CSV files which could contain embedded malicious instructions intended to influence the agent's behavior during the summary or memo-writing phase.
- Ingestion points: The
scripts/day_trading_diagnostics.pyscript reads user-provided data frominput.csvvia thepandas.read_csvfunction. - Boundary markers: No explicit boundary markers or instructions (e.g., 'ignore embedded commands') are provided to the agent when it is asked to process the data or the resulting JSON artifacts.
- Capability inventory: The skill utilizes Python scripts for file reading and writing. No network access, subprocess execution, or dynamic evaluation capabilities are present in the provided scripts.
- Sanitization: The diagnostic script performs type conversion for numeric and date fields using
pd.to_numericandpd.to_datetimewitherrors='coerce', which effectively sanitizes those specific fields. However, thesymbolfield is treated as a raw string and included in the output summary without sanitization.
Audit Metadata