wterm
Fail
Audited by Gen Agent Trust Hub on May 19, 2026
Risk Level: HIGHCOMMAND_EXECUTIONCREDENTIALS_UNSAFEREMOTE_CODE_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides scripts for spawning arbitrary shell processes on the host system via the wterm-server daemon. It also uses command-line tools like
launchctlto install and manage persistent system agents. - [CREDENTIALS_UNSAFE]: Multiple scripts (
create-session.sh,list-sessions.sh, etc.) are configured to automatically read authentication tokens from the sensitive file path~/.config/env/secrets.envfor use in API requests. - [REMOTE_CODE_EXECUTION]: The skill manages a service that exposes a terminal interface over the network (binding to
0.0.0.0:3036). This creates a persistent remote command execution interface intended for browser-based access to the host machine.
Recommendations
- AI detected serious security threats
Audit Metadata