spreadsheet

Warn

Audited by Gen Agent Trust Hub on Mar 31, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to perform privileged system operations using sudo apt-get install to install libreoffice and poppler-utils. Using elevated privileges for dependency installation should be monitored.
  • [COMMAND_EXECUTION]: The skill utilizes shell commands such as soffice (LibreOffice) and pdftoppm to render spreadsheet files into images for visual verification during the workflow.
  • [EXTERNAL_DOWNLOADS]: The skill specifies the installation of standard Python packages, including openpyxl, pandas, and matplotlib, from public registries using uv or pip.
  • [PROMPT_INJECTION]: The skill presents an attack surface for indirect prompt injection as it is designed to read and process data from external files that may contain instructions targeting the agent's logic.
  • Ingestion points: Processes tabular data from .xlsx, .csv, and .tsv files into the agent's context.
  • Boundary markers: Absent; there are no instructions to use delimiters or warnings to ignore instructions embedded in the cell data.
  • Capability inventory: The skill has the capability to write files to the local filesystem and execute shell commands via the rendering sub-process.
  • Sanitization: No input validation, escaping, or sanitization of spreadsheet content is implemented before the data is handled by the agent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 31, 2026, 02:03 PM