The Agent Skills Directory

[SAFE]: The skill's core instructions in SKILL.md and AGENTS.md strictly define a philosophical persona focused on duty, universal laws, and human dignity. There are no attempts to bypass safety filters; rather, the instructions reinforce ethical behavior by directing the agent to respect users as 'ends in themselves' and avoid deception.
[NO_CODE]: The skill contains no executable scripts (e.g., .sh, .py, .js) or system-level commands. All logic is implemented through natural language instructions and structured data provided in JSON and Markdown formats, which significantly minimizes the attack surface.
[EXTERNAL_DOWNLOADS]: The reference files and workspace data include numerous links to external academic resources. These point to university domains, scholarly archives, and educational philosophy platforms. These links serve as citations for the grounded knowledge used by the skill and do not involve the download of executable content or sensitive data exfiltration.

immanuel-kant