Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计通过

functional-specification-document功能规范文件

Agent Skill

functional-specification-document 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。

总安装

196

周安装

8

GitHub Stars

1

下载量

63
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:functional-specification-document(功能规范文件)
来源仓库:https://github.com/rahmat1929/skill-functional-specification-document
仓库路径:skills/functional-specification-document
安装命令:
npx skills add https://github.com/rahmat1929/skill-functional-specification-document --skill functional-specification-document
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/rahmat1929/skill-functional-specification-document --skill functional-specification-document

简介

functional-specification-document 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。

  • 它可结合来源仓库、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。
  • 当前顶部介绍已提供,底部简介为空,原始 SKILL.md 摘录缺失。
  • 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。

SKILL.md

Requirements & Functional Specification Document Generator

This skill produces structured Business Requirements Documents (BRDs) and implementation-ready Functional Specification Documents (FSDs). A BRD defines the business problem, goals, stakeholders, and high-level business rules. An FSD bridges the gap between what stakeholders want and what engineers need to build. It answers "what does the system do and how should it behave" without dictating internal architecture.

Content Rules

These rules govern every document this skill produces. Follow them strictly:

  • Distinguish Document Types — Know whether you are writing a BRD or an FSD. The BRD focuses on business priorities (the "why" and "what"), stricly avoiding technical constraints and implementation details. The FSD focuses on system behavior (the "how it behaves") translating those constraints to actionable scopes that developers understand.
  • NO CODE SAMPLES — Provide only explanations, conceptual descriptions, and structured information. Never include source code, pseudo-code, or implementation snippets.
  • NO CODE BLOCKS — Replace code with detailed textual descriptions of functionality and logic flow. The only permitted fenced blocks are Mermaid diagrams.
  • "What" and "why", not "how" — Use clear, developer-friendly language that focuses on system behavior and business rationale, not implementation mechanics.
  • Architectural understanding first — Prioritize system relationships, module interactions, and data flow over technical details.
  • Mermaid exclusively for visuals — Use Mermaid format for all diagrams: flowcharts, sequence diagrams, state diagrams, entity relationships, and architecture overviews. No ASCII art, no plaintext diagrams, no embedded images.
  • Exclude API & Database — The FSD does not cover API endpoint specifications or database schema design. Those belong in separate API Specification and Database Design documents respectively. The FSD references them but does not define them.

When to use this skill

  • A user wants to define the business goals and high-level requirements for an initiative (BRD)
  • A user wants to create a functional specification for a new product, feature, or system (FSD)
  • A user has a PRD, project brief, or set of requirements and needs them turned into a detailed BRD or FSD
  • A user asks to "spec out" or "write up" how something should work or what it should achieve
  • A user needs to document business goals, functional requirements, use cases, or acceptance criteria

Workflow

Phase 1: Gather Context

Before writing anything, understand what you're specifying. There are two paths depending on what the user provides:

Path A — User provides a source document (PRD, brief, requirements list, etc.): Read the document thoroughly. Extract the core requirements, identify gaps, and list clarifying questions. Present the questions to the user before proceeding. Don't guess at ambiguous requirements — ask.

Path B — No source document (interactive interview): Walk through these questions to build a mental model of the system:

  1. What is the product/feature? (one-sentence elevator pitch)
  2. Who are the users? (roles, personas, access levels)
  3. What are the core workflows? (the 3-5 things users will actually do)
  4. What external systems does this interact with? (third-party services, other internal modules)
  5. Are there constraints? (regulatory, platform, timeline, tech stack)
  6. What does success look like? (KPIs, acceptance criteria)

Don't ask all questions at once — use the first couple of answers to tailor follow-ups. The goal is to get enough information to write a solid first draft, not to exhaustively document everything upfront.

Phase 2: Write the Document

Determine if the user needs a BRD (business focus) or an FSD (system behavior focus). Read references/brd-template.md (for BRDs) or references/fsd-template.md (for FSDs) for the full document template and section-by-section guidance. Follow that template structure, but adapt it to the project — skip sections that genuinely don't apply and expand sections that need more depth.

Key principles while writing:

  • Be specific. "The system should be fast" is not a requirement. "Search results return within 200ms for up to 10,000 records" is.
  • Use consistent language. "SHALL" = mandatory, "SHOULD" = recommended, "MAY" = optional. Define this in the Document Conventions section and stick to it.
  • Every feature gets acceptance criteria. If you can't write a test for it, the requirement isn't clear enough.
  • Describe behavior, not implementation. State what the system does and why, never how it does it internally.
  • Assign priority to each requirement. Use MoSCoW: Must / Should / Could / Won't. This prevents scope creep and helps teams negotiate tradeoffs.
  • Number every requirement. Use a hierarchical ID scheme (e.g., FR-3.2.1) so requirements are traceable from spec to test to implementation.
  • Use Mermaid for all diagrams. State machines, user flows, system context, sequence diagrams — all in Mermaid format.

Phase 3: Output

Save the document as a Markdown file. Use the naming convention: BRD-[Project-Name].md or FSD-[Project-Name].md.

After generating the document, run the validation script to check structural completeness. The script validates required sections, requirement ID format, acceptance criteria presence, MoSCoW labels, Mermaid diagram presence, and flags any code blocks that shouldn't be there.

If validation reports issues, fix them before presenting the final document to the user. Show the user the validation summary alongside the finished document so they know it's been checked.

Phase 4: Review & Iterate

Present the draft to the user. Highlight:

  • Sections where you made assumptions (flag these explicitly)
  • Areas that need more detail from stakeholders
  • Requirements you marked as "Could" or "Won't" that the user might want to reconsider

Incorporate feedback and regenerate. Each revision should re-run validation.

Document Structure Summary

Business Requirements Document (BRD)

The BRD follows this top-level structure (see references/brd-template.md for details):

  1. Executive Summary — Purpose, background, and problem statement
  2. Business Goals & Objectives — Project goals and success metrics (KPIs)
  3. Project Scope — In scope and out of scope
  4. Stakeholders — Roles, influence, and responsibilities
  5. Business Requirements — High-level requirements mapped to MoSCoW priorities and business rules
  6. Assumptions, Constraints & Dependencies
  7. Glossary of Terms

Functional Specification Document (FSD)

The FSD follows this top-level structure (see references/fsd-template.md for details):

  1. Introduction — Purpose, scope, definitions, references, conventions
  2. Product Overview — Context, high-level functions, users, environment, constraints, assumptions
  3. Functional Requirements — Feature breakdown with IDs, descriptions, acceptance criteria, priority, business rules; use cases with main/alternative/exception flows
  4. User Interface Requirements — Screen descriptions, interaction behavior, navigation, accessibility
  5. Non-Functional Requirements — Performance, security, reliability, scalability, maintainability, accessibility
  6. System Behavior & Error Handling — State transitions (Mermaid), error matrix, edge cases
  7. Approval & Sign-Off — Stakeholder table, revision history
  8. Appendices — Mermaid diagrams, supplementary material

Output Quality Checklist

Before delivering the final FSD, mentally verify:

  • Every requirement has a unique ID and priority (MoSCoW)
  • (FSD only) Every feature has testable acceptance criteria
  • (FSD only) Use cases cover main flow + at least one alternative/error flow
  • (FSD only) Error handling is specified for user-facing operations
  • (FSD only) Non-functional requirements have measurable targets
  • No section is left as a placeholder or TODO
  • Cross-references between sections are consistent
  • All diagrams use Mermaid format (no code blocks, no ASCII art)
  • No code samples or code blocks appear anywhere in the document
  • No API endpoint definitions or database schemas are included
  • The document serves its audience correctly (BRD for business stakeholders, FSD for developers/QA).

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

36.92%
按下载量换算23

Claude

29.54%
按下载量换算19

Cursor

16.59%
按下载量换算10

Gemini CLI

9.59%
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。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