Token导航 LogoToken导航TokenDH.com
前端设计操作浏览器github未标认证来源可访问许可证需确认审计通过

create-regression-test创建回归测试

Agent Skill

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。

总安装

474

周安装

19

GitHub Stars

8

下载量

154
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:create-regression-test(创建回归测试)
来源仓库:https://github.com/agentmantis/test-skills
仓库路径:skills/create-regression-test
安装命令:
npx skills add https://github.com/agentmantis/test-skills --skill create-regression-test
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/agentmantis/test-skills --skill create-regression-test

简介

生成完整回归测试套件及关联测试数据。create-regression-test 属于前端设计类 Skill,可作为该场景下的辅助能力补充。

  • 适用于端到端测试覆盖率和自动化验证场景。
  • 自动创建 POM 页面对象模型和 JSON 夹具数据。
  • 安装需确认 e2e 目录结构和基础 fixtures 配置。
  • 避免为通过测试而修改真实业务逻辑,保持测试独立性。

SKILL.md

Create Regression / Smoke Test

Generate a complete test spec with associated test data and fixture integration.


Workflow

  1. Determine the feature name from the user's request
  2. Determine the suite — regression (default) or smoke
  3. Check if a POM exists in e2e/poms/{feature}.page.ts

- If yes, read its public methods and JSDoc to understand available actions - If no, create one first using the create-pom skill

  1. Check if test data exists in e2e/test-data/{feature}.json

- If yes, read and use the existing data shape - If no, create the JSON file with template data

  1. Check fixtures in e2e/fixtures/base.ts

- If the feature is not yet registered, add the interface and loader

  1. Create the spec file in the appropriate directory
  2. Run the test to verify it passes: npx playwright test --project="chromium:regression" {feature}.spec.ts

Spec File Structure

Every spec follows this exact pattern:

import { test } from '../../fixtures/base';
import { getAuthFilePath } from '../../helpers/env-config';
import { FeaturePage } from '../../poms/feature.page';

const authFile = getAuthFilePath();

test.describe('Feature Name Tests', () => {
    // One-time setup: clean stale data from previous failed runs
    test.beforeAll(async ({ browser }) => {
        const context = await browser.newContext({
            storageState: authFile,
        });
        const page = await context.newPage();
        const pom = new FeaturePage(page);
        await pom.setUp();
        await page.close();
        await context.close();
    });

    test('should create a new item', async ({ page, testData }) => {
        const pom = new FeaturePage(page);
        const item = testData.feature.items[0];

        await pom.navigateToPage();
        await pom.createItem(item.name, item.value);
        await pom.verifyItemExists(item.name);
    });

    test('should edit an existing item', async ({ page, testData }) => {
        const pom = new FeaturePage(page);
        const item = testData.feature.items[0];
        const updated = testData.feature.updatedItem;

        await pom.navigateToPage();
        await pom.createItem(item.name, item.value);
        await pom.editItem(item.name, updated.name, updated.value);
        await pom.verifyItemExists(updated.name);
    });

    test('should delete an item', async ({ page, testData }) => {
        const pom = new FeaturePage(page);
        const item = testData.feature.items[1];

        await pom.navigateToPage();
        await pom.createItem(item.name, item.value);
        await pom.deleteItem(item.name);
        await pom.verifyItemNotExists(item.name);
    });
});

Critical Rules

No Raw Page Calls in Specs

A spec must NEVER call page.getByRole(), page.locator(), page.waitForTimeout(), or any raw Playwright page method. All interaction goes through the POM.

Only acceptable uses of page in a spec:

  1. Passing it to the POM constructor: new FeaturePage(page)
  2. Creating contexts in beforeAll: browser.newContext() / context.newPage()

Test Independence

  • Every test navigates independently via pom.navigateToPage()
  • Every test creates its own data — never rely on data from another test
  • Tests run in parallel — use unique-suffixed test data to avoid collisions

Custom Test Import

Always import test from the project's custom fixtures:

// CORRECT
import { test } from '../../fixtures/base';

// WRONG — loses access to testData fixture
import { test } from '@playwright/test';

Test Data

JSON File: e2e/test-data/{feature}.json

{
    "items": [
        { "name": "ItemA", "key": "keyA", "value": "some-value" },
        { "name": "ItemB", "key": "keyB", "value": "other-value" }
    ],
    "updatedItem": { "name": "Updated", "value": "new-value" }
}

Rules

  • Never hardcode data in spec or POM files
  • All string values get unique suffixes automatically via the testData fixture
  • Shape the data to match what the POM methods expect as parameters

See references/spec-template.md and references/test-data-template.md for full templates.


beforeAll Cleanup Pattern

The beforeAll block must:

  1. Create a new browser context manually (not from the test's page fixture)
  2. Load storageState for authentication
  3. Instantiate the POM and call setUp()
  4. Close the page and context when done

This ensures stale data from previous failed runs is cleaned before any test executes.


File Checklist

When creating a regression/smoke test, you should produce:

  • e2e/poms/{feature}.page.ts — POM (create or update)
  • e2e/test-data/{feature}.json — Test data
  • e2e/fixtures/base.ts — Updated with feature interface and loader (if new feature)
  • e2e/tests/regression/{feature}.spec.ts (or tests/smoke/) — Spec file
  • All tests are independent and parallel-safe
  • No raw page calls in the spec
  • No hardcoded test data
  • Tests pass on chromium: npx playwright test --project="chromium:regression" {feature}.spec.ts

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

37.07%
按下载量换算57

Claude

30.97%
按下载量换算48

Cursor

18.42%
按下载量换算28

Gemini CLI

8.65%
按下载量换算13

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。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills