pptx

Pass

Audited by Gen Agent Trust Hub on Jun 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes system utilities through the subprocess.run interface. This includes executing soffice (LibreOffice) for document conversions, pdftoppm for image generation, and git for diffing documents. A notable technical detail in scripts/office/soffice.py is the dynamic compilation of an embedded C source string using gcc to create a shared library loaded via LD_PRELOAD; this is used to shim network sockets for compatibility in restricted execution environments.
  • [EXTERNAL_DOWNLOADS]: The skill references several external dependencies required for full functionality, including Python packages like markitdown and Pillow, and Node.js packages such as pptxgenjs, react-icons, and sharp. It also requires system-level installations of LibreOffice and Poppler utilities.
  • [PROMPT_INJECTION]: The skill is designed to ingest and parse untrusted data from external .pptx files. Although it employs defusedxml to protect against standard XML vulnerabilities, the extracted text content is interpolated into the agent's context. This creates an attack surface for indirect prompt injection, where malicious instructions embedded in a slide deck could influence the agent's future behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 15, 2026, 09:37 AM
Security Audit — agent-trust-hub — pptx