Skill flagged: https://clawhub.ai/emileindik/poku #577
Description
Skill URL: https://clawhub.ai/emileindik/poku
GitHub Repository: https://github.com/emileindik/poku
Current Version: v1.0.2
Author: emileindik
I believe our skill has been flagged incorrectly. This is a legitimate skill that allows an agent to initiate outbound phone calls on an external, sandboxed subagent. While OpenClaw is responsible for creating the prompt for the voice agent, the prompt injection concerns are exaggerated because (1) the sandboxed voice agent has no API calling abilities and can only operate on the context it is given (2) there is already plenty of precent for the use of exec with user-controlled input as a skill.
Thanks for considering, and let me know if any questions!