The Agent Skills Directory

Data Ingestion Analysis: The skill is designed to process external prompt engineering artifacts provided through the promptFiles argument. This represents a standard surface for processing user-supplied data, and the skill includes specific gates like 'behavior-test fidelity' and 'static review' to manage this interaction.
Scoped Operations: The instructions explicitly mention an 'approved write boundary' and 'HVE Builder scope approval.' These patterns are beneficial for security as they restrict the skill's ability to modify the file system or create new artifacts outside of pre-defined, authorized areas.
Verification Protocols: The workflow includes multiple validation steps, such as 'host validation' and 'quality gates.' These automated checks are used to confirm that the output of the refactoring process meets safety and functional requirements before completion.

prompt-refactor