skills/insight68/skills/pptx/Gen Agent Trust Hub

pptx

Pass

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses subprocess.run to execute soffice (LibreOffice) for converting PPTX to PDF and validating document integrity, and pdftoppm (Poppler) for generating slide images. These are standard operations for presentation processing.
  • [COMMAND_EXECUTION]: The skill uses Playwright to render local HTML slides into a headless browser for layout calculations. This execution is limited to local file paths and follows standard patterns for such automation.
  • [SAFE]: The skill utilizes defusedxml for all XML parsing operations, protecting against XML External Entity (XXE) and other XML-based vulnerabilities.
  • [PROMPT_INJECTION]: The skill ingests untrusted data from PPTX files via markitdown and unpack.py (Ingestion points: .pptx files). It lacks explicit boundary markers in its scripts (Boundary markers). It has the capability to write files and execute system commands (Capability inventory: subprocess.run for soffice, git). Sanitization is implemented via defusedxml for XML processing (Sanitization). The risk is assessed as safe/low due to the structured processing logic and use of security-hardened libraries.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 2, 2026, 03:40 PM