Token导航 LogoToken导航

Agent Skill · 研究检索

test-write(测试写入)

用于辅助测试设计、自动化测试、用例整理和回归验证。它适合让 Agent 编写单元测试、端到端测试、测试计划或根据失败日志定位问题。使用时需要确认项目测试框架、运行命令和夹具数据,避免为了通过测试而改坏真实逻辑;涉及浏览器或外部服务时,应区分本地模拟、测试环境和生产环境。

test-write 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为操作浏览器;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 69,522。

总安装 218周安装 9GitHub Stars 35下载 71核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:test-write(测试写入)
来源仓库:https://github.com/kazdenc/builder-skills
仓库路径:skills/test-write
安装命令:
npx skills add https://github.com/kazdenc/builder-skills --skill test-write
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/kazdenc/builder-skills --skill test-write

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于编写和维护各类自动化测试脚本。test-write 属于研究检索类 Skill,可作为该场景下的辅助能力补充。适合生成单元测试、API 测试或数据驱动测试用例。使用时应遵循项目编码规范与测试框架约定。确保测试代码可读性强,便于团队协作与后续维护。

安装方式:通过 GitHub 仓库安装,支持 Codex、Claude、Cursor、Gemini CLI。

来源仓库内路径:skills/test-write

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/kazdenc/builder-skills --skill test-write;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要操作浏览器,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

test-write 怎么安装?

在终端执行 npx skills add https://github.com/kazdenc/builder-skills --skill test-write。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

test-write 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

test-write 需要什么权限?

权限级别是操作浏览器。该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

test-write 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

test-write 用的人多吗?

公开数据是累计安装 218 次,近一周安装 9 次,下载 71 次,GitHub 35 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

test-write 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
test-write

测试写入 · 本页

21835操作浏览器未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Write Tests

Generate tests that catch real bugs and survive refactors. Every test should answer: what behavior breaks if this test fails?

Step 1: Read the Code

Before writing any test, understand the code under test. If the user provides a file or function, read it first.

What to identifyWhy it mattersWatch out for
InputsThese become your test parametersAssuming only happy-path inputs
Outputs / return valuesThese are your assertionsTesting internal state instead of outputs
Side effectsThese need mocking or verificationMissing async side effects, event emissions
Edge casesThese are where bugs hidenull, undefined, empty arrays, boundary values, concurrent calls
Error pathsThese need explicit test coverageOnly testing success cases

Step 2: Determine Test Type

Match the code to the right kind of test.

Code under testTest typeTools
Pure function (util, helper, transformer)Unit testVitest
React componentComponent testVitest + @testing-library/react + @testing-library/user-event
React hookHook testVitest + @testing-library/react (renderHook)
API route / server handlerIntegration testVitest + supertest or direct handler invocation
Full user flow (multi-page, auth)E2E testPlaywright
Redux / Zustand storeUnit testVitest

Step 3: Write Tests Using the AAA Pattern

Structure every test as Arrange, Act, Assert. This keeps tests readable and intention-revealing.

describe('functionName', () => {
  it('should [expected behavior] when [condition]', () => {
    // Arrange — set up inputs and dependencies
    const input = { name: 'test', value: 42 };

    // Act — call the code under test
    const result = functionName(input);

    // Assert — verify the output
    expect(result).toEqual({ processed: true, name: 'test' });
  });
});

Test Naming Convention

Use describe for the unit, it for the behavior. Name tests so a failure message reads like a sentence.

describe('calculateTotal')
  it('should return 0 when cart is empty')
  it('should sum item prices including quantity')
  it('should apply percentage discount to subtotal')
  it('should throw when discount exceeds 100%')

Common Test Patterns

Testing Async Functions

it('should resolve with user data when API call succeeds', async () => {
  const user = await fetchUser('user-123');
  expect(user).toEqual({ id: 'user-123', name: 'Test' });
});

it('should reject when user is not found', async () => {
  await expect(fetchUser('nonexistent')).rejects.toThrow('User not found');
});

Testing React Hooks

import { renderHook, act } from '@testing-library/react';

it('should increment counter', () => {
  const { result } = renderHook(() => useCounter(0));

  act(() => {
    result.current.increment();
  });

  expect(result.current.count).toBe(1);
});

Testing React Components

import { render, screen } from '@testing-library/react';
import userEvent from '@testing-library/user-event';

it('should submit form with entered values', async () => {
  const onSubmit = vi.fn();
  const user = userEvent.setup();
  render(<LoginForm onSubmit={onSubmit} />);

  await user.type(screen.getByLabelText('Email'), 'test@example.com');
  await user.type(screen.getByLabelText('Password'), 'password123');
  await user.click(screen.getByRole('button', { name: 'Log in' }));

  expect(onSubmit).toHaveBeenCalledWith({
    email: 'test@example.com',
    password: 'password123',
  });
});

Testing Components with Async State

import { render, screen, waitFor } from '@testing-library/react';

it('should display user name after loading', async () => {
  render(<UserProfile userId="123" />);

  expect(screen.getByText('Loading...')).toBeInTheDocument();

  await waitFor(() => {
    expect(screen.getByText('Jane Doe')).toBeInTheDocument();
  });
});

Testing API Routes (with mocking)

import { http, HttpResponse } from 'msw';
import { setupServer } from 'msw/node';

const server = setupServer(
  http.get('/api/users/:id', ({ params }) => {
    return HttpResponse.json({ id: params.id, name: 'Test User' });
  })
);

beforeAll(() => server.listen());
afterEach(() => server.resetHandlers());
afterAll(() => server.close());

it('should handle API error gracefully', async () => {
  server.use(
    http.get('/api/users/:id', () => {
      return new HttpResponse(null, { status: 500 });
    })
  );

  render(<UserProfile userId="123" />);

  await waitFor(() => {
    expect(screen.getByText('Failed to load user')).toBeInTheDocument();
  });
});

Testing Error Cases

it('should throw TypeError when input is null', () => {
  expect(() => processData(null)).toThrow(TypeError);
});

it('should render error boundary fallback on component error', () => {
  const BrokenComponent = () => { throw new Error('boom'); };

  render(
    <ErrorBoundary fallback={<p>Something went wrong</p>}>
      <BrokenComponent />
    </ErrorBoundary>
  );

  expect(screen.getByText('Something went wrong')).toBeInTheDocument();
});

Edge Cases to Always Consider

Include these in every test suite. They catch the bugs that slip past happy-path testing.

Edge caseExample test
null / undefined inputsprocessData(null) — should throw or return default
Empty arrays / objectscalculateTotal([]) — should return 0, not NaN
Boundary valuespaginate(items, {page: 0}) — zero, negative, max int
Concurrent callsTwo rapid clicks on submit — should not double-submit
Large inputs10,000-item array — should not crash or hang
Unicode / special characterssearch('caf\u00e9') — should handle diacritics
Type coercion trapscompare('1', 1) — should use strict equality

Code Conventions

ConventionRationale
One assertion per test (when practical)Pinpoints exactly what failed. Use multiple asserts only when they verify one logical behavior.
No test interdependencyEvery test must pass in isolation. Never rely on execution order or shared mutable state.
Use factories, not fixturescreateUser({name: 'Test'}) is flexible. fixtures/user.json is rigid and hides intent.
Colocate tests with sourceutils.ts and utils.test.ts in the same directory. Easy to find, easy to maintain.
Use vi.fn() for spiesTrack calls and arguments. Prefer over manual tracking variables.
Clean up after testsUse afterEach for DOM cleanup, timer restoration, server reset. Leaking state causes flaky tests.
Avoid test.skip accumulationSkipped tests rot. Fix or delete within one sprint.

Default Tools

ToolPurposeInstall
VitestTest runner and assertion librarynpm i -D vitest
@testing-library/reactComponent rendering and queriesnpm i -D @testing-library/react
@testing-library/user-eventRealistic user interaction simulationnpm i -D @testing-library/user-event
@testing-library/jest-domExtended DOM matchers (toBeInTheDocument, etc.)npm i -D @testing-library/jest-dom
mswNetwork-level API mockingnpm i -D msw
PlaywrightE2E browser testingnpm i -D @playwright/test

When Tests Are Done

Deliver tests that meet these checks:

  • Every test has a clear name. Reading the test name alone tells you what broke.
  • No implementation details tested. Tests don't reference internal variable names, CSS classes (prefer roles), or private methods.
  • Edge cases are covered. At minimum: null input, empty collection, error path.
  • Tests run independently. Shuffle the order — they should still pass.
  • Mocks are minimal. Only mock what crosses a boundary (network, file system, time). Let everything else run for real.

平台分布

Codex

36.01%
按下载量换算26

Claude

29.5%
按下载量换算21

Cursor

17.6%
按下载量换算12

Gemini CLI

9.2%
按下载量换算7

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills