skills/guanyang/open-agent-hub/docx/Gen Agent Trust Hub

docx

Warn

Audited by Gen Agent Trust Hub on Jul 17, 2026

Risk Level: MEDIUMCOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/office/soffice.py executes gcc via subprocess.run to compile an embedded C source string into a shared library at runtime. It then utilizes this library via the LD_PRELOAD environment variable to intercept socket system calls when executing the soffice (LibreOffice) binary. This is a form of process injection.
  • [COMMAND_EXECUTION]: The script scripts/accept_changes.py dynamically generates a LibreOffice Basic macro and writes it to a temporary application profile before executing it through the soffice command line. This allows for automated acceptance of tracked changes using internal LibreOffice functions.
  • [COMMAND_EXECUTION]: Multiple scripts throughout the skill, such as scripts/office/pack.py, ooxml/scripts/pack.py, and scripts/office/validators/redlining.py, invoke external system binaries including soffice, git, and Poppler's pdftoppm to perform document conversion, validation, and diffing.
  • [COMMAND_EXECUTION]: Documentation within SKILL.md and related files instructs the agent to perform various shell operations, including unzip, zip, and find commands, often involving files provided by external parties.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 17, 2026, 09:52 PM
Security Audit — agent-trust-hub — docx