skill-creator

Automating skill creation and evaluation involves drafting a SKILL.md, defining an eval set, and running an automated loop to test triggers and generate reports. This standardizes skill improvement cycles and reduces manual guesswork for prompt engineering.

Improving trigger accuracy requires running automated trigger evaluations and systematically refining descriptions based on feedback. This process benchmarks iterations to maximize triggering robustness across diverse prompts.

Setting up a skill iteration pipeline requires defining an initial evaluation set and executing scripts for trigger testing and reporting. This enables repeatable skill improvement and maintains consistency across multiple iterations.

Yes, you can benchmark skill iterations by running automated trigger evaluations and generating reports. This measures progress by testing triggering accuracy and robustness across a diverse set of prompts.

Yes, you need the anthropic and pyyaml dependencies to execute the automation scripts. These libraries support the underlying workflow execution and configuration parsing required for skill creation and evaluation.