office-recalculate
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external spreadsheet files, which creates a surface for indirect prompt injection if the workbook contains malicious instructions. * Ingestion points: Workbook files (
<workbook.xlsx>) processed by the LibreOffice CLI and openpyxl. * Boundary markers: No delimiters are used to separate workbook data from agent instructions. * Capability inventory: Shell command execution via Node.js and file writing to the workspace. * Sanitization: No content sanitization is mentioned for spreadsheet cells. - [COMMAND_EXECUTION]: The skill executes shell commands using Node.js and a specific CLI tool to perform formula recalculation.
- [EXTERNAL_DOWNLOADS]: The skill mentions downloading an 'Office sidecar' component. It adheres to safety standards by requiring explicit user agreement before proceeding and refers to local documentation (
sidecar.md) for the setup process.
Audit Metadata