oriental-editorial-poster

Pass

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill instructs the agent to search for and retrieve visual evidence and calligraphy samples from external web sources, such as museum collections and specialized calligraphy dictionaries (e.g., shufa5.com), which is necessary for its stated purpose.
  • [COMMAND_EXECUTION]: The skill directs the agent to save generated image outputs and prompt drafts to specific local directory paths (D:\Codex_Outputs\images and D:\Codex_Outputs\drafts). While this involves file system interaction, it is used for organizing the skill's legitimate outputs.
  • [PROMPT_INJECTION]: The skill accepts natural language inputs (keywords and themes) which are used in image generation. To mitigate potential risks, it employs a 'concept gate' that requires the agent to analyze the semantic validity of the input before proceeding with production.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 15, 2026, 04:53 PM
Security Audit — agent-trust-hub — oriental-editorial-poster