The Agent Skills Directory

[SAFE]: Network activity is restricted to established services including WeChat, Weibo, Baidu, and Toutiao for fetching trends, and various AI providers for image generation as described in the behavior declarations.
[SAFE]: The skill manages credentials using a local configuration file (config.yaml) and correctly handles access tokens for the WeChat API, following standard secret management practices.
[SAFE]: Command execution is limited to invoking the skill's own utility scripts and standard git commands for version checking and updates, all of which are aligned with the skill's functional requirements.
[SAFE]: The skill correctly handles external article content by cleaning and converting it to Markdown/HTML, with indirect prompt injection risks being low due to the intended use for stylistic learning.

zg-wewrite