← 返回 Skills 市场
kklldog

E2e Test Orchestrator

作者 Minjie.ZHOU · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ⚠ suspicious
465
总下载
0
收藏
2
当前安装
1
版本数
在 OpenClaw 中安装
/install e2e-test-orchestrator
功能描述
端到端(E2E)测试编排与执行。用于用户要求:设计测试用例、基于 Playwright/Cypress 实现自动化脚本、通过源码优先定位元素并在必要时使用截图/图像识别兜底、执行测试、自动修复脚本问题(如定位器或等待策略)、并输出结构化测试报告。
使用说明 (SKILL.md)

E2E 测试编排器

严格按以下顺序执行:测试规划 → 脚本实现 → 执行测试 → 归因/修复 → 结果报告

1)先做测试规划

  1. 阅读需求说明、验收标准、目标环境信息。
  2. 使用 references/case-template.md 产出分级测试用例:
    • P0:核心主流程(如登录、支付、鉴权、数据一致性)
    • P1:关键分支与校验流程
    • P2:边界、异常与韧性场景
  3. 每条用例必须明确:
    • 前置条件 / 测试数据
    • 操作步骤
    • 期望结果
    • 断言点与证据采集点

若需求不清晰,先写明假设,再进入脚本实现。

Playwright 关键要求(先读)

  1. 在开始实现前,先阅读 references/playwright-重点与docker兜底.md
  2. 默认尝试本地模式(local)运行 Playwright。
  3. 若本地依赖安装失败(例如 Ubuntu 25 包不兼容/缺失、无 sudo、浏览器缺库),立即切换 Docker 模式,不要阻塞流程。
  4. Docker 模式统一使用 scripts/run-playwright-docker.sh 执行,并在报告中注明镜像版本与执行模式。
  5. 可优先使用 scripts/run-playwright-auto.sh:先尝试 local,失败后自动切 docker。

2)按用例实现自动化脚本

  1. 默认优先 Playwright + TypeScript(除非用户指定其他框架)。
  2. 每条已确认用例映射为一个测试文件或一个 test(...) 用例块。
  3. 元素定位遵循 references/locator-strategy.md
  4. 使用确定性等待与断言,避免盲目 sleep
  5. 失败时必须保留证据:
    • screenshot(截图)
    • video(如已启用)
    • trace / 关键日志片段

3)执行测试并收集证据

  1. 先跑定向集(新增/变更相关用例),按需再跑全量回归。
  2. 推荐优先使用 scripts/run-playwright-auto.sh 自动选择执行模式;也可手动选择:本地用 scripts/run-playwright.sh,本地不可用时改用 scripts/run-playwright-docker.sh
  3. 记录截图、视频、trace、日志的输出路径。

4)若用例执行受阻,先归因再处理

当测试无法继续时:

  1. 先做根因分类:
    • 脚本问题(定位器错误、竞态、失效选择器、断言不正确)
    • 环境/数据问题
    • 产品缺陷
  2. 若是 脚本问题:立即修复并重跑。
  3. 若是非脚本问题:不得掩盖,按 blocker 记录复现与证据。
  4. 维护最小修复日志:
    • 失败用例 ID
    • 根因类型
    • 修复摘要
    • 重跑结果

5)输出结构化测试报告

严格按 references/report-template.md 组织输出,至少包含:

  • 测试范围与环境
  • 用例结果汇总(通过/失败/阻塞)
  • 缺陷与严重级别
  • 已实施脚本修复
  • 剩余风险
  • 证据索引(路径/链接)
  • 下一步建议

质量门槛

  • 每个自动化脚本必须可追溯到测试用例 ID。
  • 每个失败结果必须有证据支撑。
  • 阻塞用例绝不能标记为通过。
  • 对不稳定(flaky)步骤做隔离并明确记录。
安全使用建议
This skill appears to implement a normal Playwright/Cypress E2E workflow, but before installing or running it: 1) expect it to run `npm ci`/`npm i` and `npx playwright` (these will download/execute packages and may run postinstall scripts) and to pull a Playwright Docker image (default mcr.microsoft.com/playwright:v1.58.2-jammy); 2) review the repository's package.json and lockfile for untrusted packages or dangerous lifecycle scripts; 3) run in an isolated environment (CI runner or dedicated VM/container), not on a machine with sensitive credentials or production access; 4) confirm you are comfortable with Docker mounting the current working directory into the container (the scripts mount $PWD); 5) if you need to audit or pin the image, set PW_DOCKER_IMAGE to a vetted image; 6) if tests require credentials, supply them securely (not via an unreviewed agent) and avoid embedding secrets in the repo. The main technical inconsistencies are missing declared runtime dependencies (docker, node/npm/npx and optional PW_DOCKER_IMAGE) — fix or verify these before use.
功能分析
Type: OpenClaw Skill Name: e2e-test-orchestrator Version: 1.0.0 The skill bundle provides a legitimate E2E testing framework but utilizes high-risk capabilities including shell scripts and Docker execution. Specifically, `scripts/run-playwright-docker.sh` executes `docker run` with host volume mounting (`-v "$PWD":/work`) and performs automated `npm install` operations, which are high-risk actions that could be exploited for container escapes or supply chain attacks. While these behaviors are aligned with the stated purpose of running Playwright tests in isolated environments, they meet the threshold for suspicious classification due to the inherent risks of the execution model.
能力评估
Purpose & Capability
The skill claims no required binaries or env vars in its metadata, but the provided instructions and scripts expect Node/npm/npx and (optionally) Docker. Running tests will execute npm install/ci and may run code from the repository. The declared requirements do not match the actual runtime needs.
Instruction Scope
SKILL.md stays on‑topic: it instructs planning tests, implementing Playwright scripts, running tests, collecting screenshots/videos/traces, and producing reports. It directs the agent to read project source for stable selectors — appropriate for this purpose and limited to project/test artifacts.
Install Mechanism
No install spec is provided (instruction-only), but the scripts pull an official Playwright Docker image from mcr.microsoft.com and run `npm ci || npm i`. Pulling an official Microsoft Playwright image is expected; however, `npm ci` will fetch and execute packages from the npm registry (including lifecycle/postinstall scripts), which increases runtime code‑execution risk and network exposure.
Credentials
The skill declares no required env vars, yet run-playwright-docker.sh respects PW_DOCKER_IMAGE if set and scripts assume availability of docker/npx on PATH. The metadata should declare these as expected runtime dependencies. Also, executing tests typically requires test credentials/data (not requested here) — absence of guidance about secret handling is a gap.
Persistence & Privilege
The skill does not request always:true, does not modify other skills, and writes only to a local test-results directory and run.log. Autonomous invocation is enabled (default) but not combined with other high‑risk flags.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install e2e-test-orchestrator
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /e2e-test-orchestrator 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
Initial release of e2e-test-orchestrator. - Provides end-to-end (E2E) test planning, automation, execution, and reporting. - Supports Playwright (TypeScript preferred) and Cypress-based script implementation. - Includes automatic fallback to Docker execution if local dependencies fail. - Ensures tests are mapped to clear cases with structured evidence collection (screenshots, videos, traces). - Features root cause analysis, script auto-repair, and minimum fix logging. - Outputs structured, traceable test reports aligned with standard templates.
元数据
Slug e2e-test-orchestrator
版本 1.0.0
许可证 MIT-0
累计安装 2
当前安装数 2
历史版本数 1
常见问题

E2e Test Orchestrator 是什么?

端到端(E2E)测试编排与执行。用于用户要求:设计测试用例、基于 Playwright/Cypress 实现自动化脚本、通过源码优先定位元素并在必要时使用截图/图像识别兜底、执行测试、自动修复脚本问题(如定位器或等待策略)、并输出结构化测试报告。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 465 次。

如何安装 E2e Test Orchestrator?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install e2e-test-orchestrator」即可一键安装,无需额外配置。

E2e Test Orchestrator 是免费的吗?

是的,E2e Test Orchestrator 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

E2e Test Orchestrator 支持哪些平台?

E2e Test Orchestrator 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 E2e Test Orchestrator?

由 Minjie.ZHOU(@kklldog)开发并维护,当前版本 v1.0.0。

💬 留言讨论