wechat-styler

SUSPICIOUS: the formatting/conversion features fit the stated purpose, but the skill expands into authenticated browser automation that can publish/save WeChat drafts using a third-party CLI and browser profile context. Install trust looks moderate rather than overtly malicious, and data flows target the official WeChat domain, but the automation and credential/session scope are broader and riskier than a pure Markdown-to-HTML styling skill.

No clear evidence of overt malware is present. However, the module includes a high-impact capability to inject arbitrary (decoded) HTML into a live rich-text editor via bodyEditor.innerHTML with no sanitization in this module. If upstream callers can be attacker-influenced, this creates a substantial DOM XSS/content-injection risk. Verification/probing functions primarily collect state and apply image-host allowlisting, which does not mitigate the injected-markup execution risk.