Token导航 LogoToken导航TokenDH.com
开发只读github未标认证来源可访问许可证需确认审计通过

tdd时差

Agent Skill

tdd 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。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。

总安装

546

周安装

23

GitHub Stars

24

下载量

191
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:tdd(时差)
来源仓库:https://github.com/jsonlee12138/agent-team
仓库路径:skills/tdd
安装命令:
npx skills add https://github.com/jsonlee12138/agent-team --skill tdd
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/jsonlee12138/agent-team --skill tdd

简介

tdd 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合整理项目状态和变更事项。

  • 适用于围绕仓库状态、代码变更或协作事项进行信息组织和协调。
  • 通过 npx skills add 命令安装,需结合原始 README 核验具体用法。
  • 安装前建议确认权限范围和维护状态,避免触发联网或文件读写操作。
  • tdd 属于开发类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

TDD

Use this skill as a single-user workflow. Do not introduce worker, task assignment, or team coordination concepts.

Core Rules

  • Read the request, relevant code, and constraints before implementing.
  • Define what counts as done before writing production code.
  • Prefer a red phase when the task supports automated tests.
  • Do not default to E2E. Use E2E only when explicitly requested, already required by the task contract, or when lower-cost verification cannot cover the acceptance risk.
  • If the task needs test-case design, regression planning, acceptance coverage, or broader QA execution, invoke qa-expert instead of expanding this skill into a full QA workflow.
  • If automated tests are not appropriate, define manual verification before implementing.
  • Do not treat the work as complete until verification finishes.
  • If verification is skipped, state the reason and residual risk explicitly.

Workflow

Follow this sequence:

  1. Read context
  2. Define acceptance criteria
  3. Write failing tests when feasible
  4. Implement
  5. Verify
  6. Conclude

Do not skip ahead unless the previous step is clearly satisfied.

Step 1: Read Context

Inspect the request and the surrounding implementation before changing anything.

Confirm:

  • goal
  • scope
  • constraints
  • non-goals
  • existing verification options

If the request is ambiguous, stop and resolve the ambiguity before continuing.

Step 2: Define Acceptance Criteria

Write acceptance criteria before implementation.

Cover:

  • expected behavior
  • edge cases
  • failure cases
  • observable completion conditions

If acceptance criteria cannot be written yet, the task is not ready for implementation.

Step 3: Write Failing Tests When Feasible

If the task supports automated verification, write or update tests before implementation.

Target a real red phase:

  • new behavior is asserted
  • current code fails that assertion
  • expected outcome is clear

If automated tests are not a good fit, explicitly define the manual verification procedure before moving on.

Step 4: Implement

Implement against the acceptance criteria and verification plan.

Do not expand scope during implementation unless the acceptance criteria are updated first.

Step 5: Verify

Use the strongest available verification path in this order:

  1. Existing automated tests
  2. New automated tests added for the change
  3. Repeatable manual verification with concrete steps and expected results
  4. skipped only when verification is blocked in the current environment

E2E is not the default verification path. Prefer lower-cost checks first unless E2E is explicitly required or the acceptance risk cannot be covered otherwise.

If the task needs dedicated test-case design, regression execution, or broader QA coverage during verification, invoke qa-expert.

If verification fails, return to implementation and fix the issue before concluding.

Step 6: Conclude

End with exactly one of these outcomes:

  • passed
  • failed
  • skipped

When concluding, include:

  • brief summary of what changed
  • verification used
  • result
  • remaining risk if any

If the result is skipped, include the reason and the specific risk introduced by not verifying.

Failure Conditions

Stop and resolve the issue before implementation when:

  • requirements are unclear
  • acceptance criteria are missing
  • no verification strategy is defined

Loop back and keep working when:

  • tests fail unexpectedly
  • the verification command fails
  • manual verification does not satisfy acceptance criteria

Optional Artifacts

Use lightweight artifacts only when they help:

  • proposal.md
  • design.md
  • tests.md
  • explicit verify command

Do not require a fixed directory structure unless the surrounding project already has one.

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

35.05%
按下载量换算67

Claude

29.38%
按下载量换算56

Cursor

18.46%
按下载量换算35

Gemini CLI

10.14%
按下载量换算19

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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