excel-spreadsheets
Pass
Audited by Gen Agent Trust Hub on Apr 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides templates and code patterns for the agent to generate and execute Python code and VBA macros to manipulate spreadsheet data.
- [DATA_EXFILTRATION]: Contains implementation details for emailing generated Excel files using Python's
smtplib, which is a documented feature for automated reporting workflows. - [EXTERNAL_DOWNLOADS]: References several reputable and widely-used software packages for spreadsheet processing, such as
pandas,openpyxl,xlsxwriter, andexceljs. - [PROMPT_INJECTION]: The skill processes data from external Excel files, creating a surface for indirect prompt injection.
- Ingestion points: Files are read using
pd.read_excelas described inSKILL.mdandreferences/python-automation.md. - Boundary markers: The skill lacks specific instructions to the model to ignore embedded commands in the Excel data, although it does recommend structural validation.
- Capability inventory: The agent has the ability to write to the file system and perform network operations (SMTP email).
- Sanitization: Includes basic data cleaning steps like stripping whitespace and type coercion, but does not explicitly sanitize for prompt-injection style strings.
Audit Metadata