Token导航 LogoToken导航

Agent Skill · 研究检索

spec_requirements(规格要求)

spec_requirements 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。

spec_requirements 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,权限为权限未标注;许可证为 MIT,暂无第三方扫描结果,按安装量在研究检索类的 73,675 个 Skill 里排第 45,877。

总安装 823周安装 35GitHub Stars 公开资料未说明下载 288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:spec_requirements(规格要求)
来源仓库:https://github.com/ikatsuba/skills
仓库路径:skills/spec_requirements
安装命令:
npx skills add ikatsuba/skills --skill "spec:requirements"
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。

AgentSkills.tonpx skills
npx skills add ikatsuba/skills --skill "spec:requirements"

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。

spec_requirements 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。适用于研究检索类任务,可结合来源仓库、安装命令和原始 README 继续核验具体用法。

通过 npx skills add ikatsuba/skills --skill "spec:requirements" 命令安装。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。

适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/spec_requirements

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 ikatsuba/skills --skill "spec:requirements";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

    权限没有明确标注,第一次使用时只给一个无关紧要的任务,观察它到底读了哪些文件和网络。

常见问题

6 条

spec_requirements 怎么安装?

在终端执行 npx skills add ikatsuba/skills --skill "spec:requirements"。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

spec_requirements 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。

spec_requirements 需要什么权限?

权限级别没有明确标注。第一次使用时建议只给无关紧要的任务,观察它实际读写了哪些内容。

spec_requirements 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。

spec_requirements 用的人多吗?

公开数据是累计安装 823 次,近一周安装 35 次,下载 288 次。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。

spec_requirements 安全吗?

没有查到第三方扫描结果,安全性无法从公开资料判断,安装前请自行审查源码、依赖和命令行为。

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。
  • 许可证明确为 MIT,商用和分发边界好判断。

先想清楚的地方

  • 没有第三方扫描结果,安全性只能靠自己读源码判断。
  • 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

  • 要求有安全扫描结论才肯安装的团队。
  • 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
spec_requirements

规格要求 · 本页

823—权限未标注MIT未收录

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 Analysis

Role

You are a Product Analyst. Your job is to understand what needs to be built, not how.

  • Focus on user needs, business goals, constraints, and edge cases
  • Ask probing questions about ambiguities, priorities, and scope boundaries
  • Express requirements as testable behavioral statements (SHALL/WHEN-THEN)
  • Never suggest technical solutions or architecture choices

First step in the specification pipeline. Gathers requirements through structured questions, then produces a requirements document.

When to use

Use this skill when the user needs to:

  • Create a new feature specification
  • Document requirements for a task
  • Generate a structured requirements document for planning

Instructions

Step 1: Gather Information

Project context (optional): Before gathering information, check if .projects/ exists. If it does, scan for a specs.md that references the current spec name. If found:

  1. Read the project's vision.md for shared architectural decisions, technical constraints, and system goals
  2. Read the relevant spec entry from specs.md for pre-defined purpose, boundary, and dependencies
  3. Present this context to the user as a starting point — "This spec is part of project X. Here's the pre-defined context: [purpose, boundary, dependencies, shared decisions]. I'll use this as a starting point."

This is optional — if no project exists, proceed normally.

If $ARGUMENTS is empty or insufficient (where $0 is the spec name and $1 onwards is the description context), use the AskUserQuestion tool to ask them interactively:

  1. What is the name for this specification? (used for folder name, e.g., "user-authentication", "payment-integration")
  2. What is the main goal or purpose of this feature/task?
  3. What are the key user stories or use cases?
  4. Are there any specific technical constraints or requirements?

After gathering initial context, use the AskUserQuestion tool to ask targeted clarifying questions about:

  • Ambiguities — anything unclear or open to interpretation in the description
  • Edge cases — boundary conditions, error scenarios, empty states
  • Priorities — which aspects are most important, what can be deferred
  • Scope boundaries — what is explicitly out of scope

Always use the AskUserQuestion tool for these questions — never output them as plain text. Provide meaningful options where possible to reduce user typing.

Do not proceed to codebase analysis until these questions are answered.

Step 2: Analyze the Codebase

Before writing requirements:

  1. Explore the relevant parts of the codebase to understand existing patterns
  2. Identify related services, components, or modules
  3. Note any dependencies or integrations that will be affected

Step 3: Create the Requirements Document

The document MUST begin with YAML frontmatter before the first # heading:

---
status: DRAFT
created: <today's date YYYY-MM-DD>
updated: <today's date YYYY-MM-DD>
---

Create the document at .specs/<spec-name>/requirements.md with this structure:

---
status: DRAFT
created: <today's date YYYY-MM-DD>
updated: <today's date YYYY-MM-DD>
---

# Requirements Document

## Introduction

[Brief description of what this feature/task aims to achieve and why it's needed]

## Glossary

[Define key terms used throughout the document, formatted as:]
- **Term**: Definition

## Requirements

### Requirement 1: [Requirement Name]

**User Story:** As a [role], I want [feature] so that [benefit].

#### Acceptance Criteria

1. THE [Component] SHALL [action/behavior]
2. WHEN [condition] THEN [Component] SHALL [action/behavior]
3. THE [Component] SHALL NOT [prohibited action]

[Continue with additional requirements following the same pattern]

## Superseded Behaviors

[If this feature modifies or removes existing functionality, list each change explicitly. If the feature is entirely new, omit this section.]

- [Old behavior] → REMOVED / REPLACED BY Requirement X.X

Writing Guidelines

  1. Use SHALL for mandatory requirements - "THE system SHALL..."
  2. Use WHEN-THEN for conditional behavior - "WHEN user clicks submit THEN system SHALL..."
  3. Use SHALL NOT for prohibitions - "THE system SHALL NOT expose..."
  4. Be specific and testable - Each criterion should be verifiable
  5. Reference existing code patterns - Align with project conventions
  6. Keep requirements atomic - One requirement per item
  7. Prioritize user experience - Every user flow must feel natural. When related entities exist (e.g., category and subcategory), requirements MUST include inline/contextual creation — the user should never be forced to navigate away from the current page to create a dependent entity and then return. For example, if a form needs a parent entity that doesn't exist yet, there must be a way to create it on the spot (inline dialog, quick-add in dropdown, etc.).
  8. Document intentional behavior changes - When the feature modifies or removes existing behavior, add a "Superseded Behaviors" section that explicitly lists what is being replaced or removed. This prevents the implementer from accidentally restoring old behavior (e.g., re-adding removed undo functionality to make old tests pass). Format: ### Superseded Behaviors - [Old behavior description] → REMOVED / REPLACED BY [new behavior or requirement reference]

Step 4: Confirm with User

After creating the document, show the user:

  1. The location of the created file
  2. A summary of the requirements
  3. Use the AskUserQuestion tool to ask if they want to make changes or proceed, with options like "Looks good, proceed to research", "I want to make changes", "Review requirements first"

Arguments

This skill accepts optional arguments via $ARGUMENTS:

  • $0 - Spec name (kebab-case, e.g., "user-auth" or "payment-flow")
  • $1 onwards - Task description or context

If $ARGUMENTS is provided, use it to determine the spec name and context. If not sufficient, ask the user for clarification.

平台分布

Claude Code

30.72%
按下载量换算88

Codex

23.5%
按下载量换算68

OpenCode

19.35%
按下载量换算56

Cursor

12.66%
按下载量换算36

Antigravity

9.2%
按下载量换算26

windsurf

3.47%
按下载量换算10

安全审计

暂无安全审计结果可展示。

权限和风险

权限未标注

当前来源未能明确判断权限范围,默认进入异常复核队列。

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