running-skill-edd-cycle

What problem does it solve?

This Skill guides evaluation-driven development (EDD) for agent skills. Use it when setting up skill testing workflows, creating evaluation scenarios, or establishing Claude A/B feedback loops for skill validation. Provides development methodology, not content guidance.

Core Features & Use Cases

• Evaluation-first workflow: Structure tests before content to ensure skills address real user needs.
• Baseline & comparisons: Establish performance baselines and compare across iterations and models.
• Multi-model validation: Supports evaluation across multiple models and feedback loops.

Quick Start

Define 3+ evaluation scenarios, run them to establish baselines, and prepare a Claude A/B feedback plan.