Token导航 LogoToken导航TokenDH.com
开发敏感数据github未标认证来源可访问clear审计异常

requirements-engineering需求工程

Agent Skill

requirements-engineering 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

13,747

周安装

329

GitHub Stars

8

下载量

3,949
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

3

许可证

MIT

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:requirements-engineering(需求工程)
来源仓库:https://github.com/kaakati/rails-enterprise-dev
仓库路径:skills/requirements-engineering
安装命令:
npx skills add https://github.com/kaakati/rails-enterprise-dev --skill 'Requirements Engineering'
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/kaakati/rails-enterprise-dev --skill 'Requirements Engineering'

简介

处理需求文档、用户故事和验收标准的结构化整理。

  • 适合在 Codex、Claude、Cursor、Gemini CLI 中与产品团队协作。
  • 通过 GitHub 仓库安装,使用 npx skills add 命令添加技能。
  • 需区分事实需求与假设,优先引用原始需求来源。
  • 输出内容应保持可追溯性,便于后续变更追踪。

SKILL.md

Requirements Engineering

Translate natural language prompts into structured requirements and tasks.

Requirements Format Decision Tree

What format is the user using?
│
├─ "As a... I want... So that..."
│   └─ User Story format → Extract actor, feature, benefit
│
├─ "Given... When... Then..."
│   └─ BDD format → Extract precondition, action, outcome
│
├─ "Add/Implement/Build X with Y"
│   └─ Feature request → Extract action, components, technology
│
├─ "Fix/Debug X"
│   └─ Bug report → Route to debug workflow
│
└─ "Refactor/Optimize X"
    └─ Refactoring → Route to refactor workflow

NEVER Do This

NEVER accept vague requirements:

# BAD - Too vague, can't test
"Add payment saving"
"Make authentication work"

# GOOD - Specific, testable
"User can save credit card with last 4 digits displayed"
"User receives JWT token on successful email/password login"

NEVER create broad acceptance criteria:

# BAD - Not testable
- [ ] Payment should work
- [ ] Authentication is secure

# GOOD - Specific and verifiable
- [ ] Card is validated before saving
- [ ] Card number shows only last 4 digits in UI
- [ ] Access token expires in 15 minutes

NEVER skip task dependencies:

# BAD - Missing dependencies
PAY-005: Payment service (no deps listed)

# GOOD - Explicit dependencies
PAY-005: Payment service
└─ Dependencies: PAY-002 (payments table), PAY-004 (Payment model)

User Story Template

As a [actor/role]
I want [feature/capability]
So that [business benefit/value]

Example:

## User Story

**As a** customer
**I want** to save my payment method
**So that** I can checkout faster on future purchases

## Acceptance Criteria

- [ ] User can add credit card
- [ ] Card is validated before saving
- [ ] Card number masked (only last 4 digits)
- [ ] Saved cards appear in checkout dropdown

BDD Acceptance Criteria

Given [initial context/precondition]
When [action/event occurs]
Then [expected outcome/result]

Example:

**Given** a registered user
**When** they enter valid credentials
**Then** they receive a JWT access token
**And** the access token expires in 15 minutes

Task Breakdown Template

Break features into Rails technical layers:

LayerExample Tasks
DatabaseMigrations, schema changes
ModelActiveRecord, validations, associations
ServiceBusiness logic, external APIs
ControllerHTTP endpoints, routing
View/ComponentUI, Hotwire, ViewComponents
TestingRSpec tests for each layer

Dependency Flow

Database → Models → Services → Controllers → Views → Tests

Action Verb → Intent Mapping

VerbIntentWorkflow
Implement/Build/CreateNew Feature/reactree-dev
Add/IncludeEnhancement/reactree-feature
Fix/DebugBug Fix/reactree-debug
Refactor/CleanupRefactoring/reactree-refactor
Optimize/ImprovePerformance/reactree-dev --refactor

Quick Extraction Patterns

# Actor: "As a/an X"
ACTOR=$(echo "$prompt" | grep -ioE "as an? [a-z ]+" | sed 's/as an? //')

# Feature: "I want X"
FEATURE=$(echo "$prompt" | grep -ioE "i want [^.]*" | sed 's/i want //')

# Benefit: "So that X"
BENEFIT=$(echo "$prompt" | grep -ioE "so that [^.]*" | sed 's/so that //')

# Components: "with X and Y"
COMPONENTS=$(echo "$prompt" | grep -ioE "with [a-z, and]+" | sed 's/with //')

# Technology: "using Z"
TECH=$(echo "$prompt" | grep -ioE "using [a-z ]+" | sed 's/using //')

Complexity Scoring

Word CountComplexityComponentsAction
< 10Low1-2Simple task, no epic
10-20Medium2-3Feature, create tasks
> 20High4+Epic with subtasks

Beads Task Creation

For complex features, auto-create beads structure:

# Create epic
EPIC_ID=$(bd create --type epic --title "$FEATURE_TITLE")

# Create subtasks with dependencies
bd create --type task --title "Add schema" --deps "$EPIC_ID"
bd create --type task --title "Add models" --deps "$EPIC_ID,$SCHEMA_TASK"
bd create --type task --title "Add services" --deps "$MODEL_TASK"

Output Checklist

Before implementation, verify:

  • User story has actor, feature, benefit (if applicable)
  • Acceptance criteria are specific and testable
  • Technical components identified
  • Task breakdown follows layer strategy
  • Dependencies explicitly stated
  • Beads epic created (for complex features)

References

Detailed patterns and scripts in references/:

  • formats-and-patterns.md - Extraction logic, intent classification
  • task-breakdown.md - Layer strategy, dependency detection, examples
  • beads-integration.md - Auto task creation, workflow routing

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

04

需要参考平台分布和安装热度时

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

补充不同宿主或平台的使用分布数据

能力 5

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

Claude Code

29.37%
按下载量换算1,160

windsurf

22.04%
按下载量换算870

OpenCode

15.84%
按下载量换算626

Codex

11.94%
按下载量换算472

Antigravity

6.92%
按下载量换算273

Gemini CLI

3.09%
按下载量换算122

安全审计

Gen Agent Trust Hub

未通过

Socket

通过

Snyk

通过

权限和风险

敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。来源字段存在多来源差异,先按来源优先级自动处理,无法消解时进入异常复核队列。

来源信息

继续浏览同类 Skills