imagegen

Pass

Audited by Gen Agent Trust Hub on Jun 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: Detailed analysis of the skill's instructions and scripts found no malicious patterns, prompt injections, or unauthorized data access logic.
  • [COMMAND_EXECUTION]: The skill operates through provided Python scripts (scripts/image_gen.py and scripts/postprocess_image.py). These scripts are transparent and perform legitimate tasks such as calling the OpenAI API, processing JSONL batches, and performing local image manipulations using the Pillow library.
  • [EXTERNAL_DOWNLOADS]: The skill references the installation of the openai and pillow Python packages. These are well-known, industry-standard libraries from reputable sources.
  • [CREDENTIALS_UNSAFE]: Authentication is handled correctly via environment variables (LLM_API_KEY, LLM_API_ENDPOINT). The documentation refers to standard system paths for secret management rather than hardcoding sensitive information.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 17, 2026, 04:18 PM
Security Audit — agent-trust-hub — imagegen