AGENT SKILL

claude-skills-evalkit

This skill should be used when the user asks to 'test a skill', 'simulate skill triggering', 'check skill conflicts', 'run skill regression tests', 'analyze skill performance', or 'optimize a skill description'. It provides tools for testing and analyzing Claude Code skills.

使用场景测试与质量
License 许可
MIT
最近更新
2026-04-08
Stars
★ 1
Forks
0

01 前言 Front Matter

name
claude-skills-evalkit
version
0.1.0
description
This skill should be used when the user asks to 'test a skill', 'simulate skill triggering', 'check skill conflicts', 'run skill regression tests', 'analyze skill performance', or 'optimize a skill description'. It provides tools for testing and analyzing Claude Code skills.

03 SKILL.md 正文

Claude Skills EvalKit

Test and analyze Claude Code skills using the cse CLI tool.

Available Commands

Run these commands via Bash:

  • Simulate triggering: cse simulate <skill-path> -p "test prompt" — test if a skill would trigger
  • Batch test: cse simulate <skill-path> -f prompts.txt — test against multiple prompts
  • Run tests: cse test --suite <suite-path> — run regression test suites
  • Detect conflicts: cse conflicts — find overlapping skill descriptions
  • Optimize description: cse optimize <skill-path> --positive pos.txt --negative neg.txt — improve trigger precision
  • Dry run: cse dry-run <skill-path> "prompt" --diff — test instruction adherence
  • View analytics: cse dashboard — show skill usage analytics
  • Query logs: cse log query --since 7d — query session logs

Quick Start

  1. Initialize: cse init
  2. Write test prompts in a file (one per line)
  3. Run: cse simulate ./my-skill/SKILL.md -f prompts.txt -v

04 同仓库更多 Skills