The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill can retrieve terminal session history using the zmx history command, which serves as an ingestion point for untrusted data. If a terminal session displays attacker-controlled content containing hidden instructions, the agent might inadvertently execute them.
Ingestion points: Output from zmx history <name> (plain text, VT escapes, or HTML).
Boundary markers: None specified in the documentation to distinguish between terminal output and agent instructions.
Capability inventory: Terminal management commands including zmx attach, zmx run, and zmx kill which allow process manipulation.
Sanitization: No evidence of sanitization or filtering of terminal output before processing.
[Command Execution] (LOW): The skill is designed to execute arbitrary shell commands via zmx attach and zmx run. While this is the intended primary purpose of a terminal session manager, it grants the agent direct shell access, which should be monitored for unauthorized command injection.

zmx