AgentSkillsCN

test

智能测试运行器。当用户说“测试这个”、“运行测试”、“/test”或希望运行并分析测试结果时使用。它能自动识别测试框架、运行测试、解析测试结果，并针对失败情况提出修复建议。

SKILL.md
--- frontmatter
name: test
description: Intelligent test runner. Use when the user says "test this", "run tests", "/test", or wants to run and analyze test results. Detects framework, runs tests, parses results, and suggests fixes for failures.
allowed-tools:
  - Bash
  - Read
  - Glob
  - Grep

Test

智能测试运行器：检测框架 → 运行测试 → 解析结果 → 报告。

Workflow

1. Detect Framework

Check project root for test configuration:

bash
ls package.json pyproject.toml go.mod Cargo.toml Makefile 2>/dev/null
IndicatorFrameworkCommand
vitest.config.*Vitestnpx vitest run
jest.config.* / package.json has jestJestnpx jest
pyproject.toml / pytest.iniPytestpytest -xvs
go.modGo testgo test ./...
Cargo.tomlCargo testcargo test

If bun.lockb exists, prefer bun test over npx.

2. Determine Scope

  • •No arguments → run full suite
  • •File path argument → run that file only
  • •Directory argument → run tests in that dir
  • •Function/test name → run matching test (-k for pytest, -t for jest)

3. Run Tests

Run the detected command. Capture both stdout and stderr. Show the full output to the user.

4. Parse Results

Extract from output:

  • •Total tests / passed / failed / skipped
  • •Failed test names and error messages
  • •Stack traces for failures

5. Report

Format:

code
## Test Results: [pass/fail]

✓ N passed | ✗ N failed | ○ N skipped

### Failures
- **test_name**: error message
  → file:line
  → suggested fix

6. Fix Failures

If tests fail, offer to:

  1. •Read the failing test and implementation
  2. •Identify the root cause
  3. •Fix and re-run to verify

Only fix if the user agrees. Don't auto-fix without asking.