collage-layout-expert

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides high-quality technical instructions for computational collage composition. It follows best practices for modular documentation and leverages established mathematical principles.
  • [EXTERNAL_DOWNLOADS]: The skill references standard, well-known Python libraries (opencv-python, numpy, scipy, scikit-image, transformers, pot, hnswlib) for its image processing logic. These are typical for computer vision tasks.
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to execute its image processing algorithms. These algorithms are provided as transparent Python templates within the skill's own reference files, representing standard 'Code Interpreter' behavior for an expert AI agent.
  • [INDIRECT_PROMPT_INJECTION]: While the skill ingests user-provided images and text, it operates within a specialized design domain with clear technical boundaries. There is no evidence of unsafe interpolation or lack of control over the execution environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 01:25 PM
Security Audit — agent-trust-hub — collage-layout-expert