skill-coach

To create an AI agent skill following prompt engineering best practices, use structured templates and validation checklists to encode domain expertise, ensuring precise descriptions and efficient progressive disclosure of instructions.

Common anti-patterns in AI agent skill creation include vague descriptions, failing to load details progressively, and lacking structured validation. Detecting these pitfalls early ensures skills remain precise, efficient, and robust during execution.

Progressive disclosure in AI agent skills works by loading instructions efficiently and providing detailed domain expertise only when contextually needed, preventing context window bloat and maintaining optimal agent performance.

Yes, you can audit an existing AI skill for anti-patterns and description design issues by applying validation checklists that identify common pitfalls, evaluate tool usage, and verify the accurate capture of temporal knowledge.

The best way to encode domain expertise into an AI agent skill is through structured templates that translate expert knowledge into actionable instructions, paired with anti-pattern detection to ensure robust and precise agent behavior.