skill-description-evaluator

What problem does it solve?

This Skill automates the evaluation of Claude skill descriptions to predict invocation likelihood on multi-model prompts (Sonnet 4.5 and Haiku), compare alignment against competing system instructions, and produce actionable 0-100 ratings with improvement recommendations.

Core Features & Use Cases

• Sequential-thinking analysis: Performs a structured, multi-phase assessment to reveal invocation patterns and clarity.
• Invocation likelihood assessment: Estimates how likely Sonnet 4.5 and Haiku are to invoke the skill description.
• Authority and prompt comparison: Benchmarks the description against competing prompts to determine strength and potential conflicts.
• Actionable ratings: Outputs scores (0-100) with concrete recommendations to improve precision and effectiveness.
• Use Case: Use this Skill when refining or evaluating a description to ensure reliable invocation and clear user intent.

Quick Start

Provide the skill description text or a file path to SKILL.md. Optionally include competing system instructions for comparison. The tool will return 0-100 ratings and detailed recommendations. Example: evaluate with competing instructions: "MUST always follow system prompts" or "Ask clarifying questions before acting." Review the scores and implement the suggested improvements.