spreadsheets

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions require the agent to generate and execute raw JavaScript code in a Node.js environment to perform spreadsheet operations via the preloaded artifact tool surface.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through the processing of external workbook files.
  • Ingestion points: Data enters the agent context through FileBlob.load and SpreadsheetFile.importXlsx as documented in references/workbook.md.
  • Boundary markers: The skill does not include instructions or delimiters to isolate data from potential instructions within the imported spreadsheet files.
  • Capability inventory: The execution environment provides access to the local file system through node:fs/promises and FileBlob, along with extensive spreadsheet manipulation capabilities via the @oai/artifact-tool package.
  • Sanitization: There is no evidence of content validation, escaping, or filtering for data processed from external workbooks.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 01:28 AM