Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed WeChat article digest workflow with minor privacy and scoping cautions, but no evidence of deception, exfiltration, executable payloads, or unsafe persistence.
Install only if you are comfortable with the skill reading your configured public-account list, browsing public article sources, and saving a local digest in Documents. For ambiguous requests like general article summaries, confirm that you mean the WeChat/public-account digest before letting it run.
The trigger conditions are broad enough to match generic requests like '整理最近的文章' or '这周的阅读摘要' even when the user did not ask for WeChat public account processing. Over-broad activation can cause the wrong skill to run, leading to unintended browsing, reading of configured subscriptions, and downstream side effects such as local file creation.
The manifest and overview describe a tool for obtaining articles published by followed WeChat public accounts, but the workflow defines Zhihu columns/personal pages as the preferred acquisition channel and official websites as the secondary source. This means the skill is actually a cross-site article digest built from mirrors and substitutes, which can diverge from the claimed source and behavior.
The skill directs saving output to ~/Documents automatically without any user-facing warning or confirmation. Silent persistence of a digest about the user's followed accounts and reading interests can expose sensitive preference data to other local users, backups, indexing systems, or later unintended reuse.
The skill instructs the agent to write a digest file into the user's local Documents directory, but this side effect is not disclosed in the manifest description. Undisclosed local file writes can violate user expectations and create privacy or persistence risks, especially when the digest contains reading history or subscription interests.
