Security audit
Security checks for vulnerabilities and agentic risk
This skill is a straightforward WeChat article-writing helper with optional user-directed document export and no hidden code or credential access.
Before installing, users should know that generated article drafts, image-plan files, and exported Word documents may be saved locally if they choose those options. Review generated marketing or public-account content for accuracy and avoid including sensitive private information in drafts unless the workspace is trusted.
The description says the skill is invoked whenever a user needs to create a WeChat public-account article, but it does not define specific trigger phrases, scope boundaries, or exclusion conditions. This broad natural-language condition could overlap with many ordinary writing requests and cause unintended invocation in loosely related content-generation contexts.
The instruction explicitly requires that image prompts be written in Chinese, which imposes a language choice on users. There is no opt-in, alternative language option, or documented regional necessity that would justify forcing a single language.
The skill describes creating files and exporting documents as if they are saved directly, but it does not require an explicit user warning or confirmation about writing to disk at the moment of file creation. In environments with file-system tools, this can cause unintended local persistence of user content, surprising the user and potentially exposing sensitive drafts or generated material to later access.
No suspicious patterns detected.
