The Agent Skills Directory

[SAFE]: The skill consists entirely of Markdown documentation and reference material. It does not include scripts, executables, or instructions for the agent to perform dangerous system operations. The disable-model-invocation: true flag in the frontmatter further restricts the agent from making unwanted tool calls while processing this reference.
[SAFE]: External URLs provided in the source_credits and references sections point to legitimate educational resources (Notion, YouTube) and established prompt engineering guides. These are used for attribution and further reading rather than remote code execution.
[SAFE]: The 'expert personas' listed in the database are well-known historical and contemporary figures in science, business, and arts. The methodology encourages the agent to adopt these professional perspectives to improve output quality, which is a standard and safe application of Large Language Model (LLM) role-playing capabilities.

expert-domain-priming