slide-generation

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • EXTERNAL_DOWNLOADS (SAFE): The skill instructs users to install standard, well-known libraries (python-pptx, Pillow, and @marp-team/marp-cli). These are from trusted registries (PyPI, npm) and are necessary for the skill's stated purpose.
  • COMMAND_EXECUTION (LOW): The documentation demonstrates the use of CLI commands (python slides.py, npm install -g) to operate the tool. While these involve executing code locally, they are standard operational procedures for this type of utility and do not show signs of malicious command injection.
  • Indirect Prompt Injection (LOW): The skill processes structured data (slides.json) to generate content. There is a surface for indirect prompt injection if the JSON contains instructions (e.g., in slide text fields) meant to influence downstream agents or the final output.
  • Ingestion points: slides.json file input.
  • Boundary markers: None explicitly mentioned in the documentation for sanitizing JSON fields.
  • Capability inventory: Execution of slides.py which likely uses python-pptx to write files.
  • Sanitization: Not explicitly documented; the skill assumes the input JSON is well-formed for slide generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:30 PM