The Agent Skills Directory

[SAFE]: The skill implements a robust modular design, delegating specialized tasks to external skills while maintaining clear routing logic. This architecture minimizes the attack surface within this specific component.
[SAFE]: The .skillmeta.yml configuration includes proactive security measures by explicitly forbidding the upstream contribution of sensitive local files, such as .local.yml, .user/, and private/ directories.
[SAFE]: Instructions regarding the installation of external skills (e.g., hallmark, brandkit) are designed to be reported to the user as fallback commands. This ensures that any installation script execution remains under the user's explicit control and review.
[SAFE]: Analysis of the references/ directory confirms it contains standard UI/UX best practices and routing tables without any evidence of prompt injection, data exfiltration, or obfuscated payloads.

direction