excel-subtotal-audit

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from external Excel workbooks, creating a vulnerability surface where instructions embedded in cell values could attempt to influence the agent.
  • Ingestion points: The skill reads cell values and formulas from workbooks using load_workbook as described in Step 1 and the Python sketch.
  • Boundary markers: Absent. The skill does not provide specific instructions to the agent to treat cell content as data only or to ignore embedded instructions.
  • Capability inventory: The skill uses openpyxl for file reading, utilizes regular expressions for formula parsing, and prints cell content to the execution log.
  • Sanitization: Absent. Cell values are extracted and printed without escaping or validation against unexpected instruction patterns.
  • [EXTERNAL_DOWNLOADS]: The skill references the Python package openpyxl, which is a widely used and established library for spreadsheet manipulation.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 04:36 PM
Security Audit — agent-trust-hub — excel-subtotal-audit