SuperAPI AI Directory

req-tester

Testing & quality

View repository

Updated 10-15 17:34

About this skill

Requirement Testing Skill

1. Objective

Identify ambiguities, omissions, and contradictions in requirement documents, and generate a structured issue list to ensure requirements are testable.

2. Input and Output

  • Input:cleaned-requirements/index.md、CLAUDE.md
  • Output:cleaned-requirements/issues.md

3. Core Principles

  1. Testing necessity: only identify issues that affect testing work
  2. Consistency: the entire document should not contradict itself
  3. Flexibility: do not impose hard quantity limits; adapt to different scenarios
  4. Use standards for guidance, not filters and restrictions

4. Specific Standards

4.1 Issue Identification Criteria

Core judgment criterion: does this issue affect test case design or test result judgment?

If it does, it should be identified; if it does not, it does not need to be identified.

Example judgments:

  • "Does the delete operation require secondary confirmation?" → affects the testing process → identify
  • "How long is data retained?" → affects expiration scenario testing → identify
  • "What is the error prompt text?" → affects test verification → identify
  • "Why should this feature be built?" → if it does not affect test design → do not identify
  • "What database is used?" → if it does not affect test verification → do not identify
  • "What color is the button?" → if it is only decorative → do not identify; if it involves state distinction → identify

4.2 Identification Dimensions (6)

Dimension Focus
Cause and effect Business rules, permission constraints, preconditions
Boundary definition Effective conditions, scope exclusions, constraints
Exception scenarios Failure handling, retry, rollback, conflict handling
Interaction details Text, prompts, loading, empty state, feedback
Data state Field definitions, state transitions, data constraints
Performance compatibility Response time, concurrency, compatibility scope

4.3 Issue Format

Each issue contains 4 fields:

## 问题 N [分类 · 优先级]

[位置] 第X页 - 章节名称

[问题] 需求文档提到"原文引用",但没说清楚XXX,导致无法设计测试用例。

[需要澄清]
1. 具体问题1?
2. 具体问题2?

[产品回答]
<!-- 在此填写 -->

---

Field descriptions:

  • [Location]: where the issue was found
  • [Issue]: why this issue exists and what the impact is
  • [Needs clarification]: what exactly needs to be asked
  • [Product answer]: leave blank for the product team to fill in

4.4 Classification and Priority

Classification(choose 1 of 6):

  • 前因后果、边界定义、异常场景、交互细节、数据状态、性能兼容

Priority(choose 1 of 3):

  • critical: if unclear, test cases cannot be designed
  • warning: affects core test case design
  • info: affects boundary and exception test cases

5. Examples

Example 1: Correct issue

## 问题 1 [前因后果 · critical]

[位置] 第5页 - 广告账户选中

[问题] 需求文档提到"点击某一个广告账户",但没说清楚选中需要什么权限,无权限时如何处理,导致无法设计权限相关的测试用例。

[需要澄清]
1. 选中广告账户需要什么权限?
2. 没有权限时如何提示用户?

[产品回答]
<!-- 在此填写 -->

---

Why it is correct: it directly affects test case design (permission testing, exception testing)

Example 2: Incorrect issue

## 问题 X [前因后果 · warning]

[位置] 第1页 - 需求背景

[问题] 需求文档提到"通过Tiktok平台同步广告数据",但没说清楚为什么要做这个功能,市场背景是什么。

[需要澄清]
1. 为什么要开发这个功能?
2. 竞品是怎么做的?

[产品回答]
<!-- 在此填写 -->

Why it is incorrect: it is purely business background and does not affect test design

6. Checklist

  • All issues affect test case design or test result judgment
  • No purely business background issues (unless they affect testing)
  • No purely technical implementation issues (unless they affect testing)
  • No common-sense issues
  • Classification accuracy (one of 6 dimensions)
  • Priority accuracy (critical/warning/info)
  • Each issue has 4 fields: [Location], [Issue], [Needs clarification], [Product response]
  • Each issue ends with --- separator
  • Consistent wording before and after, no contradictions