skill-creator

Optimize skill descriptions by running a systematic evaluation and optimization loop that benchmarks performance across different configurations to improve triggering accuracy and ensure correct intent matching.

Creating a new skill involves capturing intent, interviewing for requirements, and drafting the SKILL.md file. You then set up test cases and use Python scripts to assist with the full lifecycle of skill development and optimization.

Yes, you can benchmark skill performance by comparing configurations and iterations. The benchmarking process uses subagents to evaluate performance, providing both quantitative metrics and qualitative feedback for optimization.

Yes, Python scripts are required to perform various tasks during skill evaluation and benchmarking. The skill optimization and testing processes rely on these scripts to execute quantitative benchmarks and qualitative feedback loops.

Test and evaluate existing skills by running quantitative benchmarks and gathering qualitative feedback. The evaluation process uses subagents to systematically test performance and identify areas for modification and optimization.