The Agent Skills Directory

[COMMAND_EXECUTION]: The batch_processor.py and delivery_packager.py scripts utilize subprocess.run to orchestrate internal logic by calling other scripts within the skill (e.g., category_detector.py, style_matcher.py). These calls use absolute paths relative to the script location and do not use a shell environment, which prevents arbitrary command injection.
[DYNAMIC_EXECUTION]: The platform_adapter.py script contains a localized call to __import__("datetime"). This is used strictly for generating ISO-formatted timestamps in the generated adaptation reports and does not process external input.
[INDIRECT_PROMPT_INJECTION]: The skill processes product information provided in products.json. While this represents an external data ingestion surface, the data is used exclusively for parameters in image processing scripts rather than as prompts for a large language model.
Ingestion points: batch_processor.py loads and parses the products.json file.
Boundary markers: No explicit boundary markers or delimiters are used for the JSON content.
Capability inventory: subprocess.run is used for internal script orchestration in batch_processor.py and delivery_packager.py.
Sanitization: The scripts use list-based arguments for subprocess calls, effectively preventing shell injection attacks, although the content of the JSON fields is not strictly validated.

ecommerce-material-studio