MUD

Operate and maintain the persistent MUD agent for OpenClaw. Use when running MUD engine commands, smoke-testing mud state behavior, validating save/restore,...

This skill appears to do what it says: it runs your local MUD engine. Before using it, inspect the local mud_agent.py (and any engine code) in the referenced workspace paths, because the helper will import or execute that file — which runs arbitrary local code. Only run this skill if you trust the engine source; prefer running it as a non-administrative user or in a sandboxed environment. Also verify there are no unexpected network calls or hardcoded secrets in the engine code if you will run commands that could trigger them. If you need platform-agnostic usage, note the README references Windows-style paths; confirm the correct workspace path on your host.

Type: OpenClaw Skill Name: mud Version: 1.0.1 The skill is classified as suspicious due to the direct execution of user-controlled input via `subprocess.run` in `scripts/mud_cmd.py`. While `shlex.split` is used to parse arguments, mitigating direct shell injection, it still presents a risk of argument injection into the underlying `mud_agent.py` script or more sophisticated prompt injection attacks against the AI agent's interpretation of the command. There is no clear evidence of intentional malicious behavior like data exfiltration or backdoor installation, but the pattern of executing user-provided commands without robust input validation beyond `shlex.split` constitutes a significant vulnerability.