evaluation

name	evaluation
description	Evaluate MCP servers and agent tools with structured test cases and metrics. Use when benchmarking tools or validation suites.

Evaluation

基于 Model Context Protocol 评估规范 | 下载时间: 2026-01-16

When to use

• User or task matches this skill's domain (see description above).
• Invoked via /skill-name or when the agent selects this skill from context.

Instructions

1. Read references/full-guide.md in this skill directory for full workflows, examples, and checklists (progressive disclosure).
2. Run scripts under scripts/ when present, using paths relative to this skill folder.
3. For Master orchestration: /master skill:evaluation when the project uses the command center.

Related

• Package root: .cursor/skills/evaluation/
• Full guide: references/full-guide.md
• Registry metadata: .cursor/features/skills/skills/registry.json