Founder-HongYun Editor Automation

This skill appears to be coherent with its stated purpose (automating a web editor plus optional WeChat publishing), but it relies on programmatic self-restraint rather than enforceable controls. Key risks: the manifest grants exec (shell) access and browser.evaluate() usage, both of which could be abused to exfiltrate cookies, credentials, or call arbitrary endpoints despite the README claiming otherwise. Before installing or using: 1) Only install from a trusted source; verify the actual published code repository if available. 2) Avoid setting persistent FOUNDER_USERNAME/FOUNDER_PASSWORD env vars—use interactive login. 3) Prefer not to provide WECHAT_APP_SECRET as a long-lived env var unless necessary; use scoped credentials and rotate/revoke after testing. 4) If possible, remove/deny exec permission or run the skill in a tightly sandboxed environment (no network egress except the platform and wechat API). 5) During first runs, monitor Browser DevTools Network tab and system process activity to ensure requests and shell calls are limited to the declared domains. 6) If you cannot audit the implementation or control exec rights, treat this skill as higher risk and consider not installing it.