Token导航 LogoToken导航TokenDH.com
前端设计external-servicegithub未标认证来源可访问许可证需确认审计通过

roblox-core罗布乐克斯核心

Agent Skill

roblox-core 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

606

周安装

25

GitHub Stars

4

下载量

198
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:roblox-core(罗布乐克斯核心)
来源仓库:https://github.com/stackfox-labs/luau-skills
仓库路径:skills/roblox-core
安装命令:
npx skills add https://github.com/stackfox-labs/luau-skills --skill roblox-core
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/stackfox-labs/luau-skills --skill roblox-core

简介

roblox-core 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息。

  • 适合在罗布乐思开发项目中围绕仓库状态进行整理。
  • 支持在 Codex、Claude、Cursor、Gemini CLI 中使用。
  • 安装前建议确认权限范围及是否触发联网或文件操作。
  • 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

roblox-core

When to Use

Use this skill when the task is primarily about core Roblox runtime structure and everyday gameplay scripting:

  • Choosing whether logic belongs on the client, server, or both.
  • Deciding where Script, LocalScript, and ModuleScript instances should live.
  • Organizing code across ServerScriptService, ServerStorage, ReplicatedStorage, ReplicatedFirst, StarterPlayer, StarterGui, and Workspace.
  • Using services, module reuse, attributes, and bindables inside normal gameplay code.
  • Working with common Studio scripting workflows like playtesting, Explorer layout, WaitForChild(), and Output-driven debugging.
  • Implementing straightforward input, camera, raycasting, collision, and CFrame behavior as part of ordinary experience scripting.

Do not use this skill when the task is mainly about:

  • Exhaustive engine API lookup or class-by-class reference browsing.
  • Cross-boundary remote design, advanced remote security, or server-authority architecture.
  • Persistence, memory stores, messaging, Open Cloud, OAuth, or external automation.

Decision Rules

  • Use this skill if the main question is structural: where code lives, what runs where, what replicates, or how to organize reusable Roblox logic.
  • Use this skill for foundational engine patterns that appear in most experiences: services, modules, attributes, bindables, basic input, workspace access, collisions, raycasts, camera, and CFrame.
  • If the task centers on RemoteEvent, RemoteFunction, trust boundaries, request validation, or multiplayer message design, hand off to roblox-networking.
  • If the task is mainly "which API/member do I call" across a large Roblox surface area, hand off to roblox-api.
  • If the task centers on saving, loading, quotas, versioning, cross-server state, or ephemeral shared state, hand off to roblox-data.
  • If the task involves Open Cloud, web APIs, credentials, OAuth, or external tooling automation, hand off to roblox-cloud or roblox-oauth.
  • If a request mixes core structure with out-of-scope systems, answer only the foundational Roblox portion and explicitly exclude the rest.
  • If unsure, prefer the narrower interpretation and omit material that would overlap networking, data, cloud, or API-reference skills.

Instructions

  1. Start by identifying the runtime side for each responsibility:

- Server for authoritative world state, spawning, rule enforcement, and shared simulation. - Client for player-local input, camera, moment-to-moment presentation, and local feedback. - Shared modules only when both sides need the same code or constants.

  1. Place code in containers that match replication behavior:

- ServerScriptService for server-only scripts and modules. - ServerStorage for server-only assets or modules that do not need to replicate. - ReplicatedStorage for shared modules and replicated assets. - ReplicatedFirst only for earliest client startup work. - StarterPlayerScripts, StarterCharacterScripts, StarterGui, and StarterPack for client behavior copied into each player.

  1. Prefer explicit script intent:

- Use LocalScript or Script with RunContext = Client for client code. - Use Script with RunContext = Server or normal server placement for server code. - Use ModuleScript for reusable logic and configuration.

  1. Retrieve services once near the top of a script with game:GetService() and keep names aligned with service names.
  2. Use WaitForChild() when accessing replicated objects from the client unless the load order is guaranteed by the container being used.
  3. Treat ModuleScript return values as cached per Luau environment:

- Require once per script and reuse the returned table or function. - Avoid circular requires. - Keep shared modules side-agnostic unless the module is intentionally server-only or client-only.

  1. Use attributes for lightweight per-instance state and configuration that should live on the instance itself.
  2. Use bindables only for communication on the same side of the client-server boundary. Prefer module-owned bindables when they simplify a local event API.
  3. For input and camera code, keep implementation client-side and adapt to the player's active input mode rather than assuming desktop-only controls.
  4. For workspace scripting:

- Read and write object state through clear references. - Use raycasts for intentional spatial queries. - Use collision groups or part properties for collision behavior. - Use CFrame operations when orientation and relative transforms matter.

  1. Keep examples and guidance at the foundational level. Do not drift into persistence, advanced networking security, or exhaustive reference lookups.

Using References

  • Open references/scripting-overview.md for the basic Roblox scripting workflow in Studio and the standard service-module-function-event script shape.
  • Open references/client-server-runtime.md to reason about authority, replication, edit versus runtime data models, and what each side can safely assume.
  • Open references/script-locations-and-script-types.md when deciding between Script, LocalScript, ModuleScript, run contexts, and container placement.
  • Open references/services.md for the core game:GetService() pattern and which container or gameplay services matter most in foundational code.
  • Open references/modulescripts-and-reuse-patterns.md for module caching, shared code placement, configuration modules, and encapsulation patterns.
  • Open references/attributes.md for per-instance state, replication-order cautions, and change-detection patterns.
  • Open references/bindable-events.md for same-side script communication, async events, sync callbacks, and argument-shape cautions.
  • Open references/input-overview.md for client-side input handling and adapting to preferred input type across devices.
  • Open references/workspace-basics-camera-raycasting-collisions-and-cframes.md for the most common world-facing runtime patterns.

Checklist

  • Each responsibility is assigned to the correct runtime side.
  • Script and module placement matches replication and visibility needs.
  • Shared code is in ModuleScript form instead of duplicated across scripts.
  • Client code uses WaitForChild() where replication order is uncertain.
  • Services are retrieved once and reused.
  • Attributes are used for lightweight instance state, not arbitrary module data.
  • Bindables are only used on one side of the client-server boundary.
  • Input and camera code stays client-side.
  • Raycasts, collisions, and CFrame operations are used intentionally for spatial logic.
  • No advanced remote-security design is included.
  • No persistence, Open Cloud, OAuth, or external API automation guidance is included.
  • No exhaustive API catalog material is included.

Common Mistakes

  • Putting server-only logic in ReplicatedStorage or other replicated containers.
  • Expecting a plain Script to run everywhere without considering location or RunContext.
  • Using LocalScript where a shared ModuleScript should hold reusable logic.
  • Assuming replicated objects already exist on the client and skipping WaitForChild().
  • Treating bindables as cross-network communication tools.
  • Mutating a module return value without realizing the cached reference is reused within that environment.
  • Using attributes for large structured data that belongs in a module or system object.
  • Driving camera or input code from the server.
  • Using Touched for non-physical overlap logic that should be a raycast or explicit spatial query.
  • Moving parts with raw position logic when relative transforms or facing direction require CFrame.

Examples

Choose placement by responsibility

-- ServerScriptService/SpawnController
-- Spawns and manages shared world state on the server.
-- StarterPlayer/StarterPlayerScripts/InputController
-- Reads player input and drives local presentation on the client.
-- ReplicatedStorage/Shared/Constants
-- Shared module used by both sides.
local Constants = {
    MaxHealth = 100,
    RoundLength = 120,
}

return Constants

Use the standard script shape

local Players = game:GetService("Players")
local ReplicatedStorage = game:GetService("ReplicatedStorage")

local RoundConfig = require(ReplicatedStorage:WaitForChild("RoundConfig"))

local function onPlayerAdded(player)
    print(player.Name, "joined; round length:", RoundConfig.RoundLength)
end

Players.PlayerAdded:Connect(onPlayerAdded)

Keep client-only camera code local

local Workspace = game:GetService("Workspace")

local camera = Workspace.CurrentCamera
camera.CameraType = Enum.CameraType.Scriptable
camera.CFrame = CFrame.lookAt(Vector3.new(0, 10, 20), Vector3.new(0, 5, 0))
camera.Focus = CFrame.new(0, 5, 0)

Use attributes and bindables for local structure

local part = script.Parent
part:SetAttribute("Active", true)

local changed = Instance.new("BindableEvent")
changed.Event:Connect(function(state)
    print("State changed:", state)
end)

changed:Fire(part:GetAttribute("Active"))

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

33.14%
按下载量换算66

Claude

31.75%
按下载量换算63

Cursor

20.72%
按下载量换算41

Gemini CLI

10.08%
按下载量换算20

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

external-service

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

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