Token导航 LogoToken导航

Agent Skill · 前端设计

init(初始化)

init 用于补充前端设计相关能力,适合在 Local Agent 中需要让 Agent 承接前端设计相关任务时使用。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。

init 是前端设计类 Agent Skill,可用宿主为 Local Agent,通过unknown获取,权限为只读;许可证信息还不明确,暂无第三方扫描结果,按安装量在前端设计类的 10,216 个 Skill 里排第 8,964。

总安装 245周安装 10下载 78核验 2026-05-01
前端设计只读unknown未标认证来源可访问许可证需确认审计未展示Local Agent

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。
许可证
unknown

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:init(初始化)
来源仓库:https://smithery.ai
仓库路径:init
安装命令:
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。当前暂无明确安装命令,请以来源页面说明为准。

这个 Skill 是做什么的

前端设计类 Skill 关注页面结构、视觉规范和组件写法,输出通常是可运行的界面代码。

init 用于补充前端设计相关能力,支持 Local Agent 参与界面初始化工作。适合处理项目脚手架搭建、依赖安装或基础样式配置。可生成标准目录结构、配置文件模板或启动脚本。安装前应检查 Node.js 环境及包管理器版本,确保兼容性。使用时应避免覆盖已有工程文件,优先采用增量更新方式。

来源仓库内路径:init

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 Local Agent;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

    没有现成安装命令,按unknown的说明从来源页面获取,仓库路径是 init。

  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

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

init 怎么安装?

公开资料没有给出可以直接复制的命令,需要按来源页面说明手动安装,仓库内路径为 init。

init 支持哪些 AI 工具?

公开资料标注支持 Local Agent。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。

init 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

init 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

init 用的人多吗?

公开数据是累计安装 245 次,近一周安装 10 次,下载 78 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。

init 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

  • 专注 Local Agent 一种宿主,触发方式和使用习惯不用来回切换。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 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 Local Agent,想直接扩展它能力的人。
  • 需要处理前端设计类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 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。

前端设计类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
frontend-design

前端设计 · Codex、Claude

8,823,360126,072只读MIT通过
vercel-composition-patterns

Vercel 组合模式 · Codex、Claude

3,858,81625,893需要联网MIT通过
vercel-react-native-skills

Vercel React native skills 前端 · Codex、Claude

2,468,66425,878只读MIT通过
adapt

adapt 搜索 · Codex、Claude

2,027,04023,125操作浏览器未明确通过
bolder

bolder 分析 · Codex、Claude

2,004,28823,130只读未明确通过
init

初始化 · 本页

245—只读未明确未收录

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

Version

Current: 1.16.0

Update this version when templates change: - Major (2.0.0): Breaking changes to file structure - Minor (1.1.0): New sections, updated rules - Patch (1.0.1): Typo fixes, clarifications

Sprint Initialization Skill

Initialize a sprint with kanban-style structure for multi-agent collaboration.


Overview

This skill creates the structure only (empty backlog). Use /sprint:add-backlog to add work items.

Generated structure:

sprints/<sprint-name>/
├── BACKLOG.md          # Work items (Feature → Task → Sub-task)
├── HANDOFF.md          # Work Board (current status)
├── INSTRUCTION.md      # Agent guidelines
├── personas/           # Agent persona files (character templates)
├── active/             # Feature-specific contexts (temporary)
└── refs/
    ├── designs/        # Feature design documents
    ├── plans/          # Task execution plans
    ├── decisions/      # Decisions (per-feature + _sprint.md for sprint-wide)
    ├── lessons/        # Lessons (per-feature + _sprint.md for sprint-wide)
    └── archive/        # Completed features (moved from BACKLOG.md)

Workflow

Phase 1: Gather Information

Ask the user:

  1. Sprint name - What should we call this sprint?

- Suggest kebab-case format (e.g., payment-system, auth-refactor)

  1. Location - Where should we create it?

- Default: sprints/<sprint-name>/ - Allow custom path if needed

  1. Team (optional) - Who will work on this?

- Persona names: #rook, #pixel, #dash, or custom names - If not specified, use default #agent

  1. Mode - How do you want to manage code?

- default - Sprint and code in same working directory (existing behavior) - worktree - Each Feature gets its own git worktree - Sprint files stay in main worktree (Single Source of Truth) - Code work happens in separate worktrees per Feature - If worktree selected: Add sprint path to .gitignore

  1. Worktree creation method (worktree mode only)

- "How should worktrees be created?" - Default: git worktree add (recommended) - Custom: User provides command template - "Where should worktrees be placed?" - Suggest: ../<project>-worktrees/ (outside main repo) - Allow custom path prefix

Phase 2: Create Structure

Create the following files from templates:

  1. BACKLOG.md from assets/templates/BACKLOG-TEMPLATE.md

- Empty backlog structure - Ready for features to be added

  1. HANDOFF.md from assets/templates/HANDOFF-TEMPLATE.md

- Work Board format - In Progress / In Review tables

  1. INSTRUCTION.md from assets/templates/INSTRUCTION-TEMPLATE.md

- Multi-agent workflow - Session start/end procedures - If team specified, add team section

  1. active/ folder

- Create empty folder - Add README explaining purpose

  1. refs/ structure

- refs/designs/ with README - refs/plans/ with README - refs/decisions/ with README + _sprint.md (sprint-wide decisions) - refs/lessons/ with README + _sprint.md (sprint-wide lessons) - refs/archive/ with README (for completed features)

  1. Version file

- Create .sprint-version with current template version (from ## Version section above) - Format: just the version number (e.g., 1.0.0)

  1. .sprint-config

- JSON config storing sprint mode and worktree settings - Created for both modes (to make mode explicit)

  1. personas/ folder Default mode: {"mode": "default"} Worktree mode: {"mode": "worktree", "sprintRoot": "<absolute-path-to-sprint-dir>", "worktree": {"command": "git worktree add {path} -b {branch}", "pathPrefix": "../<project>-worktrees"}}

- Copy all persona files from plugin assets/personas/ to sprint personas/ - These are character templates agents adopt during sessions - Users can customize or add new personas after init - worktree.command: Worktree creation command - Default: "git worktree add {path} -b {branch}" - Custom example: "./scripts/create-worktree.sh {branch} {path}" - Placeholders: {path} = worktree path, {branch} = branch name - worktree.pathPrefix: Parent path where worktrees are created - Default: "../<project>-worktrees" - Feature worktrees are created under this path

Worktree mode additional steps:

  • Add sprint path to project .gitignore
  • Include Sprint Root absolute path + Worktree Mode section in INSTRUCTION.md

Phase 3: Confirm

Show created structure and explain next steps:

Default mode:

Sprint created at: sprints/<sprint-name>/

Files:
- BACKLOG.md - Add work items with /sprint:add-backlog
- HANDOFF.md - Work board (current status)
- INSTRUCTION.md - Start work sessions with @INSTRUCTION.md #persona-name

Next steps:
1. Add work items: /sprint:add-backlog
2. Start working: @INSTRUCTION.md #rook

Worktree mode:

Sprint created at: sprints/<sprint-name>/ (worktree mode)

Files:
- BACKLOG.md - Add work items with /sprint:add-backlog
- HANDOFF.md - Work board (current status)
- INSTRUCTION.md - Start work sessions with @INSTRUCTION.md #persona-name
- .sprint-config - Sprint configuration (mode: worktree)

Sprint path added to .gitignore (sprint files are not git-tracked).

Next steps:
1. Add work items: /sprint:add-backlog
2. Start working: @INSTRUCTION.md #rook

Assets

  • assets/templates/BACKLOG-TEMPLATE.md - Backlog structure
  • assets/templates/HANDOFF-TEMPLATE.md - Work Board template
  • assets/templates/INSTRUCTION-TEMPLATE.md - Agent guidelines
  • assets/templates/FEATURE-DESIGN-TEMPLATE.md - Feature design doc
  • assets/templates/TASK-PLAN-TEMPLATE.md - Task plan doc
  • assets/templates/ACTIVE-FEATURE-TEMPLATE.md - Active feature context
  • assets/refs/designs-README.md - README for designs folder
  • assets/refs/plans-README.md - README for plans folder
  • assets/refs/decisions-README.md - README for decisions folder
  • assets/refs/lessons-README.md - README for lessons folder
  • assets/refs/archive-README.md - README for archive folder
  • assets/personas/rook.md - Pragmatist backend engineer
  • assets/personas/pixel.md - Detail-obsessed frontend engineer
  • assets/personas/dash.md - Enthusiastic fullstack prototyper
  • assets/personas/slate.md - Conservative infrastructure engineer
  • assets/personas/echo.md - Socratic discussion facilitator
  • assets/personas/thorn.md - Uncompromising code reviewer

平台分布

Local Agent

83.92%
按下载量换算65

安全审计

暂无安全审计结果可展示。

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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