Token导航 LogoToken导航

Agent Skill · 研究检索

requirements-specification(需求规格)

requirements-specification 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。

requirements-specification 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;许可证为 MIT,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 49,257。

总安装 659周安装 28GitHub Stars 175下载 231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
MIT

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:requirements-specification(需求规格)
来源仓库:https://github.com/nicepkg/ai-workflow
仓库路径:skills/requirements-specification
安装命令:
npx skills add https://github.com/nicepkg/ai-workflow --skill requirements-specification
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/nicepkg/ai-workflow --skill requirements-specification

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。

requirements-specification 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中快速定位候选结果。适用于需求规格相关的关键词搜索和信息筛选场景。通过 npx skills add 命令从 GitHub 仓库安装并使用该技能。

安装前需确认权限范围和维护状态,注意可能触发联网或文件读写操作。建议结合原始 README 核验具体用法和功能边界。

来源仓库内路径:skills/requirements-specification

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/nicepkg/ai-workflow --skill requirements-speci…;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

requirements-specification 怎么安装?

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

requirements-specification 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。

requirements-specification 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。

requirements-specification 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

requirements-specification 用的人多吗?

公开数据是累计安装 659 次,近一周安装 28 次,下载 231 次,GitHub 175 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。

requirements-specification 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。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。
  • 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

  • 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。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 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
requirements-specification

需求规格 · 本页

659175需要联网MIT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

Requirements & Specification Skill

Transform customer insights into clear, detailed specifications that engineering can build from. Master user story writing, define acceptance criteria, and manage scope ruthlessly.

User Story Writing (INVEST Format)

INVEST Principles

I - Independent (minimal dependencies) N - Negotiable (details can be discussed) V - Valuable (delivers customer value) E - Estimable (team can estimate effort) S - Small (can complete in 1-2 sprints) T - Testable (clear success criteria)

User Story Template

As a [user role]
I want [action/capability]
So that [benefit/outcome]

Acceptance Criteria:
Given [context]
When [user action]
Then [expected result]

Good vs Bad Examples

Bad Story:

As a user
I want a better dashboard
So that I can see my data

Problem: Too vague, not testable, too large

Good Story:

As a project manager
I want to see all tasks assigned to me in the last 24 hours
So that I can track what happened while I was offline

Acceptance Criteria:
Given I'm logged in
When I view the Home dashboard
Then I see a "Recent Tasks" section
And it shows tasks assigned to me from last 24 hours
And tasks are sorted by assignment time (newest first)
And clicking a task opens the task detail page

Acceptance Criteria (BDD Format)

Scenario Template

Scenario: [Specific user action]
Given [initial context/state]
When [user performs action]
Then [expected result]
And [additional verification]

Example: Password Reset Feature

Scenario: User resets password with valid email
Given I'm on the login page
And I'm not logged in
When I click "Forgot Password?"
And enter my email address
And click "Send Reset Email"
Then I see message "Check your email for reset link"
And a password reset email is sent to that address
And the email contains a valid reset link

Scenario: User uses expired reset link
Given I received a password reset email
And the reset link is more than 24 hours old
When I click the reset link
Then I see "Link has expired"
And I'm offered to request a new reset link

Scenario: Password doesn't meet requirements
Given I'm on password reset page
When I enter password "123"
Then I see error "Password must be 8+ characters"
And the form doesn't submit

Requirements Document Structure

Executive Summary (1 page)

  • Overview of feature/product
  • Business goal/context
  • Key benefits
  • Timeline
  • Success metrics

Requirements Overview (5-10 pages)

Functional Requirements

  • What the system must do
  • Features and capabilities
  • User interactions
  • Data handling

Non-Functional Requirements

  • Performance (response time < 2s)
  • Scalability (support 10K concurrent users)
  • Security (encrypt PII)
  • Availability (99.9% uptime)
  • Accessibility (WCAG AA compliance)

Business Requirements

  • Why we're building this
  • Business metrics
  • Customer need
  • Competitive advantage

Constraints

  • Technical constraints
  • Budget constraints
  • Timeline constraints
  • Resource constraints

User Stories & Epics (20-50 pages)

Structure:

  • Epic: Large initiative grouping related stories
  • User Stories: Individual features (10-20 stories per epic)
  • Tasks: Engineering breakdown (if needed)

Each Story Includes:

  • Story ID and title
  • As a... I want... so that...
  • Acceptance criteria (3-8 scenarios)
  • Story points estimate
  • Dependencies
  • Design reference (wireframe/mockup)
  • Note/clarifications

Use Cases & Flows (10-20 pages)

Use Case Template:

Use Case: [Use Case Name]
Primary Actor: [User role]
Precondition: [State before action]

Main Flow:
1. User does X
2. System responds with Y
3. User does Z
4. System returns result

Alternative Flows:
3a. If data invalid
    - System shows error
    - User corrects and resubmits

Data Models (10 pages)

Entity Relationship Diagram

  • Entities (User, Post, Comment)
  • Relationships (User creates Posts)
  • Attributes (Post title, content, creation_date)
  • Primary keys, foreign keys

UI/Wireframes (Attached)

  • User interface mockups
  • User flows and navigation
  • Key interactions

Scope Management

MVP vs Nice-to-Have

MoSCoW Method:

MUST Have (Non-negotiable)

  • Core functionality
  • Without these: product won't work
  • Must launch with these
  • Example: User login, basic content view

SHOULD Have (Important but not critical)

  • Enhance user experience
  • Value add
  • If time allows
  • Example: Advanced search, saved preferences

COULD Have (Nice-to-have)

  • Polish features
  • Low priority
  • Do if extra time/budget
  • Example: Dark mode, animations

WON'T Have (Explicitly out of scope)

  • Clear for future
  • Helps say "no" to stakeholders
  • Plan for later version
  • Example: Mobile app (launching web first)

Scope Creep Prevention

Red Flags:

  • "Can we just add...?"
  • "This would be better if..."
  • "What about also including..."
  • "One more thing..."

Responses:

  • "That's a great idea. Let's add it to the roadmap for Q2."
  • "That would add 3 weeks. What would you deprioritize?"
  • "That's outside current scope. Document for next phase."

Change Management

Change Request Process:

  1. Document the change
  2. Assess impact (time, complexity)
  3. Present trade-offs
  4. Get stakeholder decision
  5. Update requirements document
  6. Communicate to team

Common Pitfalls

Too Vague

❌ "Improve performance" ✅ "Reduce page load time from 4s to under 2s"

No Success Criteria

❌ "Build dashboard" ✅ "Build dashboard showing active users in last 24h with 95% accuracy"

Missing Context

❌ "Fix the bug" ✅ "When searching with special characters, results show error. Fix to handle special chars."

Over-Specifying

❌ "Use Redux with saga middleware for state management" ✅ "State changes must be traceable and debuggable"

Ambiguous Acceptance Criteria

❌ "System should be fast" ✅ "API response time < 200ms for 95th percentile"

Requirements Review Checklist

  • ✓ Each requirement is testable
  • ✓ No requirement specifies implementation
  • ✓ Dependencies identified and documented
  • ✓ Acceptance criteria clear and complete
  • ✓ Engineering has estimated effort
  • ✓ Design mockups provided
  • ✓ Data models documented
  • ✓ Edge cases considered
  • ✓ Scope clearly defined (MVP vs future)
  • ✓ Success metrics identified
  • ✓ Timeline realistic
  • ✓ Reviewed by engineering lead
  • ✓ Reviewed by design lead
  • ✓ Stakeholder aligned

Troubleshooting

Yaygın Hatalar & Çözümler

HataOlası SebepÇözüm
Story çok büyükEpic olarak yazıldıStory breakdown
AC belirsizVague criteriaGiven/When/Then format
Scope creepChange mgmt yokChange request process
Missing edge casesHappy path focusEdge case workshop

Debug Checklist

[ ] Her story INVEST criteria geçiyor mu?
[ ] Acceptance criteria testable mı?
[ ] Non-functional requirements tanımlı mı?
[ ] Dependencies documented mı?
[ ] Engineering review yapıldı mı?

Recovery Procedures

  1. Ambiguous Requirements → Clarification meeting
  2. Scope Creep → Trade-off matrix
  3. Missing Feasibility → Engineering spike

Write clear requirements and avoid 90% of project problems!

平台分布

OpenCode

29.27%
按下载量换算68

Cursor

24.66%
按下载量换算57

Claude Code

18.76%
按下载量换算43

github-copilot

12.7%
按下载量换算29

Gemini CLI

7.62%
按下载量换算18

cline

3.47%
按下载量换算8

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。

继续浏览同类 Skills