The Agent Skills Directory

[SAFE]: The skill implements logic to guide the agent's behavior by analyzing user intent and routing requests to appropriate specialized tools.
The instructions define a clear decision-making flow without attempting to override system constraints or safety filters.
No evidence of data exfiltration, credential theft, or persistence mechanisms was found.
[EXTERNAL_DOWNLOADS]: The skill includes documentation and instructions for installing additional modules from the vendor's repository.
Evidence: npx -y skills add liustack/liustack --skill ls-xxx in SKILL.md.
These references point to the verified vendor's own repository ('liustack') and are provided as manual commands for the user, which is a standard procedure for modular toolsets.