etsy-tony-full-photo-15
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONPROMPT_INJECTION
Full Analysis
- [DYNAMIC_EXECUTION]: The skill utilizes a local Python script located at
scripts/color_cast_check.pyto measure and correct color casts in generated images. Analysis of the script confirms it uses standard libraries (numpy,Pillow) for image processing and performs only local file I/O operations without network access. - [COMMAND_EXECUTION]: The instructions provide specific shell commands for the agent to execute the color analysis script (e.g.,
python scripts/color_cast_check.py <dir> --json ...). This execution is a core, documented feature of the skill's quality control process. - [PROMPT_INJECTION]: The
SKILL.mdfile contains a 'Mandatory prompt' section that the agent is instructed to use verbatim. This prompt establishes a specific role-play persona ('资深的etsy产品设计师') and set of constraints for the image generation task. While it uses strong instructional language, it does not attempt to bypass safety guidelines or override system-level security protocols.
Audit Metadata