The Agent Skills Directory

[COMMAND_EXECUTION]: Executes a variety of system diagnostic commands including process discovery (ps), network socket listing (lsof), and routing table inspection (route).
[COMMAND_EXECUTION]: Pipes data from a local Unix socket (/var/tmp/verge/verge-mihomo.sock) to a Python interpreter. This pattern is used to parse connection metadata, rules, and configurations from a local proxy service.
[COMMAND_EXECUTION]: Directly interpolates user-provided arguments into shell commands such as 'nc', 'route', and 'ssh' without explicit sanitization, which could lead to command injection if malicious arguments are provided.
[CREDENTIALS_UNSAFE]: Accesses sensitive RustDesk configuration files (RustDesk2.toml, RustDesk_local.toml) and log files which may contain connection IDs, settings, or other private metadata.
[CREDENTIALS_UNSAFE]: Attempts to read SSH public keys (/root/id_ed25519.pub) on the remote server as part of the diagnostic routine.
[REMOTE_CODE_EXECUTION]: Executes a multi-line diagnostic script on a remote server via SSH. While intended for server troubleshooting, this involves running code with root privileges on a target defined by user input.

rustdesk-doctor