Documents

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [Persistence Mechanisms] (LOW): The Xlsx/recalc.py script modifies the local LibreOffice configuration by writing a StarBasic macro to the user's application profile (e.g., ~/.config/libreoffice/4/user/basic/Standard/Module1.xba). This modification persists across sessions to enable Excel formula recalculation.
  • [Dynamic Execution] (LOW): Xlsx/recalc.py dynamically generates and saves StarBasic code at runtime before executing it through the LibreOffice binary. This behavior is tied to the skill's primary function.
  • [Command Execution] (LOW): The skill uses subprocess.run to invoke the soffice (LibreOffice) binary for document validation and formula recalculation. While functional, this creates a dependency on external binary execution.
  • [Indirect Prompt Injection] (LOW): The skill processes untrusted documents (PDF, DOCX, PPTX, XLSX) and extracts data into structured formats interpreted by the agent. 1. Ingestion points: PDF and Office files provided by users. 2. Boundary markers: Instructional guidelines are present in Pdf/forms.md, but extracted data lacks strict delimiters. 3. Capability inventory: Includes subprocess execution of LibreOffice, file system writes, and XML parsing. 4. Sanitization: Employs the defusedxml library to mitigate XML External Entity (XXE) vulnerabilities during file parsing.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 04:34 PM