Token导航 LogoToken导航

Agent Skill · 前端设计

acceptance-testing(验收测试)

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。

acceptance-testing 是前端设计类 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,按安装量在前端设计类的 10,216 个 Skill 里排第 5,161。

总安装 727周安装 30GitHub Stars 1下载 238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:acceptance-testing(验收测试)
来源仓库:https://github.com/pixel-process-ug/superkit-agents
仓库路径:skills/acceptance-testing
安装命令:
npx skills add https://github.com/pixel-process-ug/superkit-agents --skill acceptance-testing
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/pixel-process-ug/superkit-agents --skill acceptance-testing

这个 Skill 是做什么的

前端设计类 Skill 关注页面结构、视觉规范和组件写法,输出通常是可运行的界面代码。

用于辅助编写单元测试、端到端测试及测试计划,确保功能实现与规格一致。适用于需要自动化回归验证、用例整理或根据日志定位问题的开发环境,支持多种测试框架。通过将验收标准转化为测试用例,建立从规范到代码的验证链条,防止提前宣称完成。

安装命令:npx skills add https://github.com/pixel-process-ug/superkit-agents --skill acceptance-testing 需确认项目测试配置、运行命令及夹具路径,避免为通过测试而破坏真实逻辑或引入副作用。

来源仓库内路径:skills/acceptance-testing

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/pixel-process-ug/superkit-agents --skill accep…;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

acceptance-testing 怎么安装?

在终端执行 npx skills add https://github.com/pixel-process-ug/superkit-agents --skill acceptance-test…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

acceptance-testing 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。

acceptance-testing 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。

acceptance-testing 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。

acceptance-testing 用的人多吗?

公开数据是累计安装 727 次,近一周安装 30 次,下载 238 次,GitHub 1 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。

acceptance-testing 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。
  • 需要处理前端设计类任务、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。

前端设计类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
frontend-design

前端设计 · Codex、Claude

8,823,360126,072只读MIT通过
vercel-composition-patterns

Vercel 组合模式 · Codex、Claude

3,858,81625,893需要联网MIT通过
vercel-react-native-skills

Vercel React native skills 前端 · Codex、Claude

2,468,66425,878只读MIT通过
adapt

adapt 搜索 · Codex、Claude

2,027,04023,125操作浏览器未明确通过
bolder

bolder 分析 · Codex、Claude

2,004,28823,130只读未明确通过
acceptance-testing

验收测试 · 本页

7271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

Acceptance Testing

Overview

Acceptance-driven backpressure connects specification acceptance criteria directly to test requirements, creating a validation chain that prevents premature completion claims. The system cannot cheat — you cannot claim a feature is done unless tests derived from spec acceptance criteria actually pass.

Announce at start: "I'm using the acceptance-testing skill to validate against specification criteria."


The Backpressure Chain

+------------+     derives      +------------+     validates    +------------+
|   SPECS    |---------------->|   TESTS    |---------------->|   CODE     |
|            |                  |            |                  |            |
| Acceptance |                  | Test cases |                  | Must pass  |
| Criteria   |                  | from AC    |                  | all tests  |
+------------+                  +------------+                  +------------+
      ^                                                              |
      |                    backpressure                               |
      +--------------------------------------------------------------+
      If tests fail, implementation must change (not the spec or test)

Phase 1: Extract Acceptance Criteria

Goal: From each specification file, extract all Given/When/Then acceptance criteria.

Actions

  1. Locate all specification files (specs/*.md)
  2. Extract every acceptance criterion with its ID
  3. Document in structured format

Example Extraction

## From spec: 01-color-extraction.md

### AC-1: Extract dominant colors
- Given an uploaded image (PNG, JPG, or WebP)
- When color extraction is triggered
- Then 5-10 dominant colors are returned
- And each color includes hex, RGB, and HSL representations

### AC-2: Handle invalid images
- Given a corrupted or unsupported file
- When color extraction is attempted
- Then an appropriate error is returned
- And no partial results are produced

STOP — HARD-GATE: Do NOT proceed to Phase 2 until:

  • All spec files are located and read
  • Every acceptance criterion is extracted with an ID
  • Criteria are in Given/When/Then format
  • No criteria are ambiguous (if ambiguous, clarify with spec author)

Phase 2: Derive Test Cases

Goal: Map every acceptance criterion to at least one test case.

┌─────────────────────────────────────────────────────────────────┐
│  HARD-GATE: Every acceptance criterion must have at least one   │
│  corresponding test. No exceptions. If a criterion has no       │
│  test, the feature is NOT complete.                             │
└─────────────────────────────────────────────────────────────────┘

Traceability Table

Acceptance CriterionTest TypeTest DescriptionTest File:Line
AC-1: Extract dominant colorsIntegrationUpload valid image, verify 5-10 colors with hex/RGB/HSLtest/color.test.js:15
AC-2: Handle invalid imagesIntegrationUpload corrupted file, verify error, verify no partial datatest/color.test.js:42

Decision Table: Test Type for Acceptance Criteria

Criterion TypeTest TypeRationale
Data input/output behaviorIntegrationTests real data flow
Error handling behaviorIntegrationTests error paths end-to-end
Performance requirementLoad testRequires measurement under load
UI behaviorE2E (Playwright)Tests real browser interaction
Subjective qualityLLM-as-judgeCannot be deterministically tested
Security requirementIntegration + security testTests authorization and input validation

STOP — HARD-GATE: Do NOT proceed to Phase 3 until:

  • Every acceptance criterion has at least one test mapped
  • Test types are appropriate for the criterion type
  • Test file locations are identified

Phase 3: Write Tests Before Implementation

Goal: Write acceptance tests that will fail until the feature is correctly implemented.

Actions

This phase integrates with test-driven-development:

  1. Write test from acceptance criterion (RED)
  2. Implement feature to pass test (GREEN)
  3. Refactor while keeping test green (REFACTOR)

Behavioral Outcome Focus

Verify This (Behavioral)NOT This (Implementation)
"5-10 colors are returned""K-means runs with k=8"
"Response time < 200ms""Cache is hit on second call"
"Error message is user-friendly""CustomError class is thrown"
"Data persists across sessions""PostgreSQL INSERT executes"
"UI updates within 500ms""WebSocket message is received"

STOP — HARD-GATE: Do NOT proceed to Phase 4 until:

  • All acceptance tests are written
  • Tests fail before implementation (RED confirmed)
  • Tests verify behavioral outcomes, not implementation details

Phase 4: Validation Gates

Goal: Before claiming any task complete, ALL gates must pass.

GateCheckToolRequired
Unit testsAll passTest runnerAlways
Integration testsAll passTest runnerAlways
Acceptance testsAll AC-derived tests passTest runnerAlways
BuildCompiles without errorsBuild toolAlways
LintNo violationsLinterAlways
TypecheckNo type errorsType checkerWhen applicable
┌─────────────────────────────────────────────────────────────────┐
│  HARD-GATE: ACCEPTANCE                                         │
│                                                                 │
│  Cannot claim completion without ALL acceptance tests passing.  │
│  If any acceptance test fails, the feature is NOT done.        │
│  Fix the implementation, not the spec or the test.             │
└─────────────────────────────────────────────────────────────────┘

STOP — HARD-GATE: Do NOT proceed to Phase 5 until:

  • All validation gates pass
  • Acceptance tests pass with green status
  • No gates are skipped or marked as "will fix later"

Phase 5: Traceability Report

Goal: Produce a report linking every spec criterion to its test and result.

Report Template

## Acceptance Test Report

| Spec | Criterion | Test | Status |
|------|-----------|------|--------|
| 01-color-extraction.md | AC-1: Extract dominant colors | test/color.test.js:15 | PASS |
| 01-color-extraction.md | AC-2: Handle invalid images | test/color.test.js:42 | PASS |
| 02-palette-rendering.md | AC-1: Render palette grid | test/palette.test.js:8 | PASS |

### Summary
- Total criteria: N
- Tested: N
- Passing: N
- Failing: 0
- Coverage: 100%

Anti-Patterns / Common Mistakes

Anti-PatternWhy It Is WrongCorrect Approach
Changing specs to match implementationDefeats the purpose of specificationFix the implementation, not the spec
Skipping edge case criteriaEdge cases cause production bugsALL acceptance criteria get tests
Testing implementation detailsBrittle tests that break on refactorTest observable behavioral outcomes
Claiming "tests pass" without acceptance testsUnit tests alone are insufficientAcceptance tests are a separate, required category
Writing acceptance tests after implementationTests shaped to pass, not to specifyWrite BEFORE implementation (TDD)
Deferring acceptance tests to "later"Later never comesWrite them in Phase 2, before coding
Marking failing tests as "known issues"Hides incomplete implementationFix the code until tests pass

Rationalization Prevention

ExcuseReality
"The unit tests cover this"Unit tests test components in isolation; acceptance tests verify integrated behavior
"The spec is obvious, no need for formal tests"Obvious specs still need verifiable tests
"We can manually verify this"Manual verification is not repeatable or trustworthy
"The acceptance criteria are too vague to test"Clarify the criteria; vague specs produce vague code
"This is just a cosmetic change"Cosmetic changes can break layout, accessibility, and UX

Integration Points

SkillRelationship
spec-writingAcceptance criteria come from specs
test-driven-developmentTDD cycle uses acceptance-derived tests
llm-as-judgeFor subjective criteria that cannot be deterministically tested
verification-before-completionFinal verification includes acceptance test check
autonomous-loopExit gate requires acceptance tests passing
code-reviewReview checks acceptance test coverage
planningPlan includes acceptance test writing as explicit tasks

Skill Type

RIGID — The backpressure chain must not be bypassed. Every acceptance criterion must have a test. No completion without passing acceptance tests. Fix the implementation, not the spec or the test.

平台分布

Codex

33.71%
按下载量换算80

Claude

30.94%
按下载量换算74

Cursor

19.09%
按下载量换算45

Gemini CLI

8.51%
按下载量换算20

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