skill-forge

You can manage the AI skill lifecycle by creating, auditing, improving, and reverse-engineering skills. This process ensures quality control and efficiency by providing a structured rubric, overlap detection, and batch audits to prevent common skill management issues.

You can batch-audit AI skills in a specific directory to detect issues like missing descriptions, excessive token counts, and duplicate triggers. Run the audit command targeting your skill directory path to automatically identify and report these common quality problems.

The best way to reverse-engineer an AI skill is through prompt reversal functionality. This feature analyzes existing skills to extract their underlying structure and logic, allowing you to understand, evaluate, and reconstruct the original prompt design.

Yes, you can detect overlapping triggers across multiple AI skills. The skill creation process includes built-in overlap detection, which identifies duplicate or conflicting triggers to ensure each skill operates independently and prevent common management issues.

To analyze and rewrite an AI skill for better performance, use the skill improvement functionality. It evaluates existing skills to suggest targeted improvements, including structural rewrites and performance enhancements, ensuring higher quality and operational efficiency.

You need a scoring rubric for AI skill creation when you want to ensure structured quality control. The rubric provides a standardized evaluation process for new skills, helping to prevent common issues and maintain high efficiency during skill management.