code-slides

Pass

Audited by Gen Agent Trust Hub on May 14, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The scripts/generate_image_deck.py script fetches content from Google's official Generative Language API (generativelanguage.googleapis.com) to generate slide images. This is a well-known and trusted service.
  • [COMMAND_EXECUTION]: The skill includes automation scripts for deck scaffolding (scripts/scaffold_deck.py) and skill installation (scripts/install_local.sh). These perform expected file system operations and directory management to integrate the skill into the user's environment.
  • [PROMPT_INJECTION]: The skill processes external markdown and JSON content specifications through scripts/generate_image_deck.py to create prompts for AI image generation. This represents an indirect prompt injection surface.
  • Ingestion points: Markdown or JSON files provided via the --content argument in scripts/generate_image_deck.py.
  • Boundary markers: The script parses sections but does not implement specific 'ignore' instructions for embedded content.
  • Capability inventory: The script performs network POST requests to the Gemini API and writes image files to the local disk.
  • Sanitization: Implements basic slugification for filenames but passes content directly into prompt templates.
Audit Metadata
Risk Level
SAFE
Analyzed
May 14, 2026, 02:34 PM
Security Audit — agent-trust-hub — code-slides