finance-tracker
Pass
Audited by Gen Agent Trust Hub on Sep 10, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process external financial data, such as historical performance records and budget tables, which creates a surface for indirect prompt injection if the ingested data contains malicious instructions.
- Ingestion points: Data is ingested through the
historical_dataparameter in Python classes and thefinancial_datatable referenced in SQL queries. - Boundary markers: The instructions mandate a 'Financial Accuracy First' approach requiring validation of all data sources, though specific structural delimiters or 'ignore' instructions for data content are not provided in the templates.
- Capability inventory: The skill performs SQL analysis and uses Python for cash flow forecasting and investment modeling (NPV, IRR).
- Sanitization: The provided code templates focus on mathematical logic and do not implement explicit input sanitization or filtering for non-numeric instructional content within the input data.
Audit Metadata