xlsx-processing-anthropic

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [Dynamic Execution] (MEDIUM): The script 'recalc.py' dynamically generates a LibreOffice Basic macro (Module1.xba) and writes it to the user's application configuration directory (~/.config/libreoffice or ~/Library/Application Support/LibreOffice). While this is a functional requirement for the skill to perform recalculations, writing to config paths and executing generated code via subprocess is a sensitive behavior.
  • [Indirect Prompt Injection] (LOW): The skill possesses a data ingestion surface for external Excel files and has the capability to write to the filesystem and execute system commands.
  • Ingestion points: 'recalc.py' uses the 'openpyxl' library to read user-provided Excel files.
  • Boundary markers: None; cell content is processed directly to identify error strings.
  • Capability inventory: The script uses 'subprocess.run' to execute 'soffice' and 'timeout' commands and performs file writes via the LibreOffice macro.
  • Sanitization: Cell data is not sanitized or escaped before being included in the JSON result summary returned to the agent.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 01:39 PM