The Agent Skills Directory

[SAFE]: The skill consists entirely of natural language instructions formatted as markdown. These instructions guide the AI to adopt a terse, efficient communication style, often referred to as 'Honey'. There are no indications of malicious intent.
[NO_CODE]: No scripts, executables, or third-party package dependencies are present in the skill files. The skill operates purely within the context of the AI's system prompt or project instructions.
[PROMPT_INJECTION]: While the instructions command the AI to 'Apply reflexively; never spend reasoning on how to comply,' this is interpreted as a stylistic guideline for brevity. The skill explicitly instructs the AI to NEVER cut warnings, risks, or safety-relevant caveats regarding health, money, legal issues, or irreversible actions, showing alignment with safety protocols.

honey-chat