skills/phuongnamsoft/skills/pptx/Gen Agent Trust Hub

pptx

Warn

Audited by Gen Agent Trust Hub on Apr 28, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes system tools and performs runtime code compilation.\n
  • Files scripts/thumbnail.py, scripts/office/soffice.py, and scripts/office/validators/redlining.py use subprocess.run to call soffice, pdftoppm, git, and gcc.\n
  • scripts/office/soffice.py dynamically generates C code, compiles it into a shared object, and uses LD_PRELOAD to inject the shim into the soffice process environment.\n- [EXTERNAL_DOWNLOADS]: The skill requires multiple third-party dependencies to be installed from official registries.\n
  • Dependencies include markitdown, Pillow, pptxgenjs, react-icons, react, react-dom, and sharp.\n- [PROMPT_INJECTION]: The skill processes untrusted Office documents, which serves as a potential surface for indirect prompt injection.\n
  • Ingestion points: scripts/office/unpack.py and scripts/thumbnail.py ingest user-supplied slide files.\n
  • Boundary markers: Absent in instructions for data interpolation into agent context.\n
  • Capability inventory: Significant capability surface including subprocess execution and file system write access.\n
  • Sanitization: The skill properly uses defusedxml to mitigate XML-based vulnerabilities during parsing.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Apr 28, 2026, 05:08 PM
Security Audit — agent-trust-hub — pptx