Token导航 LogoToken导航TokenDH.com
开发敏感数据github未标认证来源可访问许可证需确认审计提醒

spec规格说明

Agent Skill

spec 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

250

周安装

10

GitHub Stars

42

下载量

81
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:spec(规格说明)
来源仓库:https://github.com/taylorhuston/local-life-manager
仓库路径:skills/spec
安装命令:
npx skills add https://github.com/taylorhuston/local-life-manager --skill spec
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/taylorhuston/local-life-manager --skill spec

简介

规格说明用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息。

  • 适合在 Codex、Claude、Cursor、Gemini CLI 中围绕仓库状态、代码变更或协作事项进行整理。
  • 通过 npx skills add 命令从指定 GitHub 仓库安装,需结合原始 README 确认具体用法。
  • 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操作。
  • 该工具适用于需要自动化处理代码协作流程的场景,提升开发效率。

SKILL.md

/spec

Manage protocol-level specifications - the contract defining *what* a system must do.

What is a Spec?

A spec (specification) defines requirements at the protocol/standard level:

  • External specs: Standards you implement (LEAF spec, OAuth, OpenAPI)
  • Self-authored specs: Your own protocol defining what your system does

Specs are NOT feature breakdowns or epics. They are the source of truth for requirements.

Usage

/spec                           # Show current project's spec status
/spec --import <url>            # Import external spec (GitHub, raw URL)
/spec --init                    # Create new protocol spec for project
/spec --sync                    # Sync imported spec with upstream
/spec --section <name>          # Show specific section of spec

File Structure

spaces/[project]/
├── docs/
│   ├── specs/                  # The protocol spec (source of truth)
│   │   ├── README.md           # Spec overview and compliance status
│   │   ├── api-specification.md # API contract
│   │   ├── data-models.md      # Data structures
│   │   ├── required-features.md # Feature requirements
│   │   └── ...
│   └── adrs/                   # Architecture decisions
└── src/                        # Implementation

ideas/[project]/
├── project-brief.md            # Strategy (private)
└── issues/
    └── 001-auth/
        └── TASK.md             # implements: docs/specs/required-features.md#authentication

Why specs live with code:

  • Specs are the contract the code fulfills
  • Developers need them alongside implementation
  • Changes to spec and code can be atomic commits
  • All documentation (specs, ADRs) lives together in docs/

Execution Flow

1. Determine Context

Read: ideas/[project]/project-brief.md    # Strategy context
Glob: spaces/[project]/docs/specs/*.md    # Existing specs

Questions:

  • Does this project implement an external spec?
  • Or does it need its own protocol spec?

2a. Import External Spec

For projects implementing a standard (like leaf-nextjs-convex → LEAF spec):

/spec --import https://github.com/leafspec/spec

Process:

  1. Clone/fetch spec files
  2. Copy to spaces/[project]/docs/specs/
  3. Create docs/specs/README.md with:

- Source URL and version - Last synced date - Compliance checklist

  1. Suggest initial TASKs based on spec sections

Sync upstream changes:

/spec --sync

2b. Create Protocol Spec

For projects that need their own spec (like coordinatr):

/spec --init

Conversational creation:

  1. What does this system do? (elevator pitch)
  2. Who are the actors/users?
  3. What are the core operations?
  4. What are the API boundaries?
  5. What are the data models?

Output structure:

# [Project] Specification

## Overview
[What this system does and why]

## Actors
[Who/what interacts with the system]

## Core Operations
[The fundamental things the system must do]

## API Specification
[Endpoints, inputs, outputs, errors]

## Data Models
[Entity definitions, relationships, constraints]

## Required Features
[Feature requirements organized by domain]

## Test Criteria
[How to verify compliance]

3. Spec Status Dashboard

/spec  # No arguments

Shows:

  • Spec source (external URL or self-authored)
  • Last updated/synced
  • Sections and their implementation status
  • Linked TASKs per section

Spec vs Old "Feature Specs"

Old Model (Wrong)New Model (Correct)
SPEC-001, SPEC-002...Single protocol spec
Feature breakdownRequirements contract
Internal planning docsSource of truth
Created per featureCreated once, evolved
TASKs link to SPEC-###TASKs implement spec sections

Integration with /issue

When creating a TASK, link to the spec section it implements:

---
implements: docs/specs/required-features.md#authentication
---

The /issue command will prompt:

"Which spec section does this implement? (or 'none' for standalone)"

Compliance Tracking

Status is tracked inline within spec documents at the requirement level:

### §1 Authentication

**Requirements:**
- ✅ User registration with email/password
- ✅ User login with JWT token
- ⏳ Password reset flow
- ⏳ Email verification

**API Endpoints:**
- ✅ `POST /api/auth/register`
- ✅ `POST /api/auth/login`
- ⏳ `POST /api/auth/reset-password`

Status markers:

  • ✅ Implemented and working
  • 🚧 In progress
  • ⏳ Not started

This allows granular visibility into what's done without referencing private TASKs.

The /complete command updates these markers when work is finished.

Self-Authored Spec Guidelines

When creating your own protocol spec:

  1. Be specific - Vague specs lead to vague implementations
  2. Define boundaries - What's in scope vs out of scope
  3. Include test criteria - How do you verify compliance?
  4. Version it - Specs evolve; track changes
  5. Keep it stable - Changes should be deliberate

Spec Versioning (differs from ADRs)

ADRs are immutable - changes create a new superseding document.

Specs are edited in place - they're living contracts that evolve:

  1. Frontmatter version - Use semantic versioning (version: 1.0.0)
  2. Git history - Preserves full evolution
  3. Git tags - Mark release points (git tag spec-v1.0.0)
  4. CHANGELOG - Note significant spec changes

Version bumps:

  • Patch (1.0.1): Typos, clarifications, no behavior change
  • Minor (1.1.0): New optional features, backwards compatible
  • Major (2.0.0): Breaking changes, removed requirements

This keeps specs simple while git provides the audit trail.

Workflow

/spec --init or --import    # Define what to build
        ↓
/issue                      # Create work items that implement spec sections
        ↓
/plan                       # Break down implementation
        ↓
/implement                  # Build against the spec
        ↓
/complete                   # Verify spec compliance

Related Commands

  • /issue - Create TASKs that implement spec sections
  • /plan - Break down implementation of a TASK
  • /validate-spec - Check implementation against spec
  • /project-status - See spec compliance overview

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

36.93%
按下载量换算30

Claude

28.53%
按下载量换算23

Cursor

19.1%
按下载量换算15

Gemini CLI

10.36%
按下载量换算8

安全审计

Gen Agent Trust Hub

可疑

Socket

通过

Snyk

可疑

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。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