Excel Analysis
Pass
Audited by Gen Agent Trust Hub on Apr 5, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill uses well-established Python libraries (pandas, matplotlib, openpyxl) for data analysis tasks. All code examples are standard practices for reading, processing, and writing Excel files.
- [PROMPT_INJECTION]: The skill involves reading data from external sources (Excel files), which is a characteristic of indirect prompt injection surfaces. However, this is the primary and intended function of the skill, and no malicious exploitation patterns were observed.
- Ingestion points: The skill reads external data using
pd.read_excel()andpd.ExcelFile()from various spreadsheet files mentioned inSKILL.md. - Boundary markers: The templates do not currently include explicit instructions for the agent to treat spreadsheet content as untrusted data.
- Capability inventory: The skill includes capabilities to write data to local files using
to_excel()andsavefig(). - Sanitization: The provided examples do not include data validation or sanitization, which is typical for general-purpose analysis templates.
Audit Metadata