Wcs Helper Network Skill

SSH tunnel for China servers to access internationally blocked sites (GitHub, ClawHub, HuggingFace, arXiv, Google, YouTube). Password-auth based, one-command...

This skill may be trying to do what it claims (create an SSH SOCKS5 tunnel), but there are several red flags you should address before installing: - Missing scripts: The package only includes tunnel_handler.py, but SKILL.md and README repeatedly refer to connect.sh and many helper scripts that are not present. Do not run commands that download or execute missing scripts from unknown URLs without inspecting them first. Ask the author for the full package or the missing files and inspect connect.sh and any systemd install scripts. - Password handling: The instructions tell you to send your server password in a chat command. That exposes the password to the messaging service, agent processing, and any logs. Prefer SSH key authentication and avoid sending secrets in chat. If you must use a password, verify where and how it is stored (inspect ~/.wcs_tunnel.conf) and ensure it is not sent to external endpoints. - sshpass risks: The skill recommends sshpass, which places the password on the command line and can be observed by other processes/users. Consider using key-based auth or an SSH agent instead. - systemd auto-start: The skill mentions installing a systemd service. Only allow that after reviewing the exact service file to confirm it does nothing unexpected. - If you still want to try it: run in an isolated VM/container, create a throwaway VPS account, replace password auth with a dedicated limited account, and audit the connect.sh and related scripts before giving any real credentials. If the author cannot provide the missing scripts for review, avoid installing.

Type: OpenClaw Skill Name: wcs-helper-network-skill Version: 1.0.1 The skill facilitates setting up an SSH tunnel to bypass network restrictions but introduces significant security risks by instructing users to provide VPS credentials, including passwords, via plaintext slash commands in SKILL.md. This practice could expose sensitive credentials in chat logs or history. Additionally, scripts/tunnel_handler.py uses subprocess.run(shell=True), which presents a shell injection risk if inputs are not properly sanitized. While these behaviors are highly insecure, they appear to be functional requirements for the stated purpose rather than intentional malice.