expense-report
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions are focused on legitimate financial data processing and report generation.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided receipt data, which is its primary function.
- Ingestion points: Receipt details provided via screenshots, photos, or text (SKILL.md).
- Boundary markers: None explicitly defined in the instructions.
- Capability inventory: Generates Excel workbooks and uses spreadsheet formulas (SKILL.md).
- Sanitization: Instructions mandate extracting specific structured fields (Date, Category, Amount, etc.), which acts as a natural filter for unstructured input.
Audit Metadata