shijing-paper-zine

Pass

Audited by Gen Agent Trust Hub on Aug 31, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides detailed instructions for an AI agent to act as a visual designer. It uses local assets (assets/*.png) for style calibration and does not attempt to download external scripts or execute remote code.
  • [SAFE]: There are no signs of credential exposure or sensitive file access. The skill manages user input through a structured 'Scene Note' process which limits the potential for direct prompt injection.
  • [SAFE]: No obfuscation techniques such as Base64, zero-width characters, or homoglyphs were found in the provided files.
  • [SAFE]: The instruction set includes a 'Quality Gate' and 'Scene Routing' to ensure the output remains consistent with the requested artistic style, showing no signs of persistence mechanisms or privilege escalation.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 31, 2026, 09:48 AM
Security Audit — agent-trust-hub — shijing-paper-zine