postiz

The skill is coherent with its stated social-posting purpose and uses a plausible official install path, so it is not malicious. Risk is still high because it grants an AI agent the ability to publicly post/manage content across many connected accounts, upload local media, and optionally redirect API traffic via a custom endpoint.

This module implements standard OAuth2 device-flow login/logout/status and stores tokens locally with restrictive permissions, with no clear malware behaviors such as persistence, data corruption, or covert exfiltration. However, it contains a meaningful client-side security risk: openBrowser() uses child_process.exec with a shell command string that directly interpolates a URL derived from a remote auth-server response (verification_uri), creating potential OS command injection. Additionally, authServer/apiUrl are configurable without allowlisting, increasing risk if misconfigured or attacker-influenced.