The Agent Skills Directory

[SAFE]: The skill consists of Markdown, JSON, and YAML files without any executable code. This design minimizes security risks associated with script execution.\n- [NO_CODE]: No Python, JavaScript, or Shell scripts are included in the package. The skill operates entirely through instructions to the AI model.\n- [SAFE]: No hardcoded credentials, data exfiltration logic, or unauthorized network requests were found in any of the 29 files.\n- [SAFE]: The skill includes transparent instructions and mechanisms such as 'Protection Lists' (保護清單) to ensure that sensitive data like prices and legal commitments are preserved verbatim during editing.

humanizer-zh