image-generation

Pass

Audited by Gen Agent Trust Hub on Jun 18, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill integrates with established services (Google Gemini/Imagen and OpenAI DALL-E) and utilizes official, well-maintained libraries from PyPI (google-genai, openai, pillow, requests).
  • [COMMAND_EXECUTION]: Uses the Bash tool as intended to execute Python scripts that perform API calls and image processing tasks.
  • [CREDENTIALS_UNSAFE]: Correctly instructs users to manage API keys via environment variables, avoiding the risk of hardcoded secrets.
  • [PROMPT_INJECTION]: Instructions focus on legitimate prompt engineering techniques to improve image quality without attempting to override the agent's core safety guidelines or operational constraints.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 18, 2026, 05:27 PM
Security Audit — agent-trust-hub — image-generation