The Agent Skills Directory

[SAFE]: The skill is a standard wrapper for the official wuji command-line utility provided by the vendor. It facilitates hardware visualization without introducing malicious patterns.
[DATA_EXPOSURE]: The skill documents the --lan and --bind flags, which allow exposing the Rerun visualization server to the local network. The documentation explicitly warns the user that this configuration lacks authentication and TLS, recommending its use only on trusted networks.
[PROMPT_INJECTION]: The instructions include safety constraints for the AI agent, such as requiring explicit user confirmation before performing tactile zeroing. These instructions are designed to prevent accidental hardware miscalibration rather than bypassing safety filters.
[COMMAND_EXECUTION]: The skill uses the wuji binary as a prerequisite. This is expected behavior for a tool-integration skill and does not involve arbitrary command execution or privilege escalation.

wuji-cli-viz