The Agent Skills Directory

[SAFE]: The skill is primarily instructional, defining a three-step internal reasoning process (Self, Mirror, Response) to improve the conversational tone of the agent.
[NO_CODE]: The included Python script in the reference guide is a benign string parser using standard library functions for processing text output. It does not perform any file system or network operations.
[PROMPT_INJECTION]: While the skill includes instructions to change the AI's behavior based on user triggers (e.g., "be real", "less AI"), these are stylistic directives and do not attempt to bypass safety guidelines or override administrative constraints.

behuman