The Agent Skills Directory

[SAFE]: The skill defines a set of natural language instructions to modify the AI's output format (abbreviations, fragment usage). It does not contain any malicious patterns such as credential harvesting, unauthorized network access, or command execution.
[PROMPT_INJECTION]: While the skill includes instructions to maintain a persistent state ('No reverting or drift until explicitly stopped'), these instructions are scoped to the communication style and explicitly allow for exceptions such as security warnings and irreversible action confirmations, indicating adherence to safety standards.

caveman