excel
Pass
Audited by Gen Agent Trust Hub on Apr 9, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill includes instructions to download and install the
uvpackage manager from its official domainastral.shusing piped shell commands.astral.shis the established service for this widely-used developer tool. - [COMMAND_EXECUTION]: The provided scripts (
excel-readerandexcel-writer) execute shell commands using theuvenvironment manager to run Python logic for processing spreadsheet files. - [PROMPT_INJECTION]: The skill acts as a surface for indirect prompt injection by ingesting untrusted data from Excel files and presenting it to the agent for analysis. This is a common property of tools designed for file data extraction.
- Ingestion points:
packages/excel/main.pyreads data from user-supplied Excel files (.xlsx,.xlsm) using theopenpyxllibrary. - Boundary markers: The skill currently lacks explicit delimiters or specific instructions for the agent to ignore embedded commands within the processed data.
- Capability inventory: The skill possesses the ability to read from and write to the local filesystem via its utility scripts.
- Sanitization: Data from spreadsheet cells is converted to standard formats (Markdown and JSON) for the agent without additional filtering of the cell content.
Audit Metadata