The Agent Skills Directory

[SAFE]: The skill implements a secure methodology for extracting configuration data. The script manifest_parser.py uses a custom, restricted grammar parser to statically analyze theme manifests, avoiding the use of eval() or other dangerous JavaScript execution methods when reading data literals from .js files.
[SAFE]: All identified Python dependencies (numpy, Pillow) are well-known, industry-standard libraries used for image processing and numerical analysis. The installation instructions provided in measure_preview.py are standard and do not involve piped remote code execution.
[SAFE]: The workflow and scripts focus on local file operations, such as scaffolding templates (scaffold_template.py) and measuring rendered image properties (measure_preview.py). These activities are consistent with the skill's primary purpose and show no signs of unauthorized network access or data exfiltration.
[SAFE]: The interactive components in workbench.js use standard browser APIs (such as navigator.clipboard and Blob for JSON export) to manage configuration data locally within the user's browser environment.

liuguang-banlan-ui