Token导航 LogoToken导航

Agent Skill · 研究检索

sdd-tdd-implement(SDD TDD 实施)

sdd-tdd-implement 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

sdd-tdd-implement 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,权限为需要联网;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 72,863。

总安装 186周安装 8GitHub Stars 20下载 65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:sdd-tdd-implement(SDD TDD 实施)
来源仓库:https://github.com/sivaprasadreddy/sdd-skills
仓库路径:skills/sdd-tdd-implement
安装命令:
npx skills add https://github.com/sivaprasadreddy/sdd-skills --skill sdd-tdd-implement
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/sivaprasadreddy/sdd-skills --skill sdd-tdd-implement

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。

sdd-tdd-implement 用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词、任务场景或来源线索快速定位候选结果。通过 npx skills add 命令从指定 GitHub 仓库安装,需结合原始 README 确认具体用法。

安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。注意该技能当前分类为研究检索,实际功能以来源仓库文档为准。

来源仓库内路径:skills/sdd-tdd-implement

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/sivaprasadreddy/sdd-skills --skill sdd-tdd-imp…;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

    它需要需要联网,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

sdd-tdd-implement 怎么安装?

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

sdd-tdd-implement 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。

sdd-tdd-implement 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

sdd-tdd-implement 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。

sdd-tdd-implement 用的人多吗?

公开数据是累计安装 186 次,近一周安装 8 次,下载 65 次,GitHub 20 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。

sdd-tdd-implement 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
sdd-tdd-implement

SDD TDD 实施 · 本页

18620需要联网未明确通过

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

SDD: TDD Implementation

You are a senior software engineer executing the implementation plan using strict Test Driven Development.

Core TDD discipline: For every unit of behaviour, you write a failing test first, then write the minimum production code to make it pass, then refactor. You never write production code without a failing test that demands it.

The cycle is: 🔴 Red → 🟢 Green → 🔵 Refactor. Announce each phase explicitly as you work.


Pre-conditions

Verify these files exist before starting:

  • plan.md — the implementation plan
  • feature.md — the feature spec (acceptance criteria drive the TDD cycle)
  • docs/project.md — project context and conventions

If any are missing, stop and tell the user which file is absent and which skill produces it.


Process

1. Read Everything First

Read plan.md, feature.md, and docs/project.md in full before writing a single line of code.

From feature.md, extract the full acceptance criteria list. This is your TDD backlog — every AC must be driven by a failing test before any production code for it is written.

From docs/project.md, note:

  • Testing libraries in use (JUnit 5, Mockito, Testcontainers, RestAssured, etc.)
  • Architecture pattern — this determines what kind of tests to write at each layer
  • Package naming and REST base path conventions

2. Establish the Foundation (No TDD Required)

Before the TDD cycle begins, set up structural scaffolding that does not contain behaviour. These items do not require a failing test first because they contain no logic to test:

  • Database migration files (V<n>__<description>.sql)
  • JPA entity classes (fields, annotations, no business logic)
  • Repository interfaces (Spring Data method signatures only)
  • DTO / record definitions
  • Package structure and empty class shells

Create these first, then run:

mvn compile   (or ./gradlew compileJava)

Fix any compile errors before proceeding. Do not move to the TDD cycle until the project compiles cleanly.

Announce when scaffolding is complete:

✅ Scaffolding complete — project compiles cleanly.
Starting TDD cycle for AC-01.

3. TDD Cycle — One Acceptance Criterion at a Time

Work through the AC list from feature.md sequentially. For each AC, complete the full Red → Green → Refactor cycle before starting the next.


🔴 RED — Write a Failing Test

Announce:

🔴 RED — AC-<n>: <AC description>
Writing failing test: <TestClass#methodName>

Rules for the Red phase:

  • Write one test that directly asserts the behaviour described in the AC.
  • The test must fail for the right reason — because the production behaviour does not exist yet, not because of a compile error or missing import.
  • Choose the correct test type for the layer being tested: Layer Test type Annotations / tools Domain / service logic Unit test — pure Java, no Spring context @ExtendWith(MockitoExtension.class) Repository queries Slice test @DataJpaTest + Testcontainers REST controller Slice test @WebMvcTest + MockMvc Full request → DB flow Integration test @SpringBootTest + RestAssured + Testcontainers
  • Name the test method to describe the scenario: should_<expectedOutcome>_when_<condition> (e.g., should_throwException_when_emailAlreadyExists)
  • Do not write any production code during this phase.

Run the test and confirm it fails:

mvn test -Dtest=<TestClass#methodName>   (or ./gradlew test --tests "<TestClass.methodName>")

Show the failure output. If it passes without production code, the test is wrong — fix it before proceeding.


🟢 GREEN — Write the Minimum Production Code

Announce:

🟢 GREEN — Writing minimum production code to pass: <TestClass#methodName>

Rules for the Green phase:

  • Write only the code required to make the failing test pass.
  • Resist the urge to generalise, handle edge cases not covered by the test, or refactor.
  • It is acceptable (and expected) for this code to be imperfect — that is what Refactor is for.
  • If multiple implementation files are needed (e.g., service + repository), create them now, but keep each method minimal.
  • Do not write new tests during this phase.

Run the test again and confirm it passes:

mvn test -Dtest=<TestClass#methodName>

Then run the full test suite to confirm nothing regressed:

mvn test   (or ./gradlew test)

If any previously passing test now fails, fix the regression before proceeding. Do not carry failures forward.


🔵 REFACTOR — Improve Without Changing Behaviour

Announce:

🔵 REFACTOR — Cleaning up after AC-<n>

With all tests passing, now improve the code:

In production code, look for:

  • Duplication that can be extracted to a private method or shared utility
  • Magic literals that should be named constants or @ConfigurationProperties
  • Method or class names that do not clearly express intent
  • Violation of the architecture conventions in docs/project.md (e.g., business logic in a controller)
  • Missing Javadoc on new public API methods
  • Overly complex conditionals that can be simplified

In test code, look for:

  • Duplicated setup that belongs in @BeforeEach
  • Test data builders or factory methods that could reduce boilerplate across tests
  • Assertion messages that would be clearer on failure

After refactoring, run the full test suite again to confirm all tests still pass:

mvn test   (or ./gradlew test)

Report what was refactored (or "No refactoring needed" if the Green code was already clean).


Proceed to Next AC

Once Red → Green → Refactor is complete and all tests pass, announce:

✅ AC-<n> complete. Starting AC-<n+1>.

Repeat the cycle for every remaining AC.


4. Cross-Cutting Concerns (After All ACs Are Green)

After every AC has a passing test and the code is refactored, address concerns that span multiple ACs:

  • Input validation: Add Bean Validation annotations (@Valid, @NotNull, @Size, etc.) to request DTOs. Write a test per validation rule.
  • Error handling: Confirm the GlobalExceptionHandler (or equivalent) maps domain exceptions to the correct HTTP status codes. Write tests for each error path.
  • Security: Confirm that endpoints requiring authentication are protected. If using Spring Security, test with @WithMockUser or equivalent.
  • Logging: Add log statements at appropriate levels (INFO for business events, WARN/ERROR for failures). Logging does not require a failing test — add after Green.
  • OpenAPI annotations: Add @Operation, @ApiResponse etc. to controller methods if the project uses Springdoc.

Run the full test suite after adding each concern.


5. Final Acceptance Criteria Verification

Once all ACs are complete, do a final sweep:

For each AC in feature.md:

  • Identify the primary test covering it
  • Run that test in isolation to confirm it still passes
  • Check the AC checkbox in feature.md: change - [] to - [x]

Run the complete test suite one final time:

mvn test   (or ./gradlew test)

Do NOT declare the feature done if any AC is unchecked or any test is failing.


6. TDD Summary Report

Produce the completion summary:

## TDD Implementation Complete

### TDD Cycle Summary
| AC    | Test Class & Method                        | Red ✓ | Green ✓ | Refactor ✓ |
|-------|--------------------------------------------|-------|---------|------------|
| AC-01 | FooServiceTest#should_..._when_...         |  ✓    |   ✓     |     ✓      |
| AC-02 | FooControllerTest#should_..._when_...      |  ✓    |   ✓     |     ✓      |

### Files Created
- `src/main/java/...` — description
- `src/test/java/...` — description

### Files Modified
- `src/main/java/...` — description

### Test Suite
- Total tests: N
- Passing: N
- Failing: 0

### Notes
Any deviations from the plan, design decisions made during TDD, or
emergent behaviour discovered through the tests.

Then prompt the user to run /sdd-review before archiving.


TDD Rules — Never Break These

  1. No production code without a red test. If you're writing production code and there is no failing test demanding it, stop and write the test first.
  2. One failing test at a time. Do not write multiple failing tests before making them pass. Complete the full cycle for each before moving on.
  3. Minimum code to pass. In the Green phase, resist over-engineering. The Refactor phase exists for a reason.
  4. Never break existing tests. Run the full suite after every Green and after every Refactor. Fix regressions immediately — never carry them forward.
  5. Tests must fail for the right reason. A test that fails with a compile error is not a red test — it is a broken test. Fix the compile issue before counting it as Red.
  6. Test names are documentation. A future reader must understand exactly what scenario is covered from the method name alone, without reading the test body.

平台分布

Codex

36.21%
按下载量换算24

Claude

31.74%
按下载量换算21

Cursor

20.38%
按下载量换算13

Gemini CLI

9.22%
按下载量换算6

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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