The Agent Skills Directory

[SAFE]: The skill consists entirely of natural language instructions for agent behavior management. It contains no executable code, network operations, or external dependencies.
[COMMAND_EXECUTION]: The protocol explicitly mandates a full stop of all file edits, code changes, and task execution when triggered. This functions as a safety mechanism to prevent unintended actions during high-stress interactions.
[PROMPT_INJECTION]: The 'Escape Hatch' section allows a user to bypass the listening protocol with specific phrases like 'just do it'. This is a design feature for user autonomy and does not attempt to override platform safety filters or system-level constraints.

calm-down