@tank/skill-creator

You create AI agent skills by synthesizing domain expertise from books and web research, structuring the knowledge into reference files, and writing a SKILL.md to activate the agent. This process covers scoping, research, planning, writing, and deployment.

Evaluating AI agent skill quality involves creating test cases, grading the outputs, and iterating on the skill design. This framework provides structured methodologies to test, grade, and improve skill performance progressively.

Structuring knowledge synthesis for a prompt engineering skill involves extracting best practices from authoritative sources, organizing them into reference files and assets, and applying writing conventions to ensure progressive disclosure.

Yes, you can build domain-specific AI skills without prior experience because the workflow guides you through scoping, research, planning, and writing. It transforms general-purpose agents into domain specialists using structured reference files.

Progressive disclosure in AI skill creation is a writing convention that controls how information is revealed to the agent contextually. It is needed when structuring deep domain expertise into reference files to avoid overwhelming the agent.

Skill creation works directly with existing research workflows by leveraging books and web research to build deep domain expertise. It synthesizes this external knowledge into structured assets and SKILL.md files for deployment.