The Agent Skills Directory

[SAFE]: The skill purely consists of natural language instructions to modify the agent's reasoning process and output style. It does not invoke any tools or perform system operations.
[SAFE]: Explicit safety guardrails are integrated into the instructions, specifically requiring the agent to bypass prose compression and use full clarity when issuing security warnings or confirming irreversible/destructive actions.
[SAFE]: No evidence of obfuscation, remote code execution, or credential harvesting was found. The use of non-Latin characters (Classical Chinese) is contextually appropriate for the described stylistic output levels and does not constitute a homoglyph attack.

sharp-coder