Token导航 LogoToken导航

Agent Skill · 前端设计

game-feel(游戏感觉)

game-feel 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。

game-feel 是前端设计类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为执行命令;许可证信息还不明确,第三方扫描已通过,按安装量在前端设计类的 10,216 个 Skill 里排第 6,472。

总安装 499周安装 20GitHub Stars 1下载 162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
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:game-feel(游戏感觉)
来源仓库:https://github.com/jarrodmedrano/jarrod-claude-skills
仓库路径:skills/game-feel
安装命令:
npx skills add https://github.com/jarrodmedrano/jarrod-claude-skills --skill game-feel
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/jarrodmedrano/jarrod-claude-skills --skill game-feel

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。

用于处理 GitHub 仓库、Issue 和代码协作信息。适合在 Codex、Claude 等宿主中围绕代码变更进行整理。可结合来源仓库和 README 进一步核验具体功能。安装前建议确认权限范围和维护状态。game-feel 属于前端设计类 Skill,可作为该场景下的辅助能力补充。

注意检查是否会触发命令执行或文件读写操作。

来源仓库内路径:skills/game-feel

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 https://github.com/jarrodmedrano/jarrod-claude-skills --skill gam…;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

    它需要执行命令,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

game-feel 怎么安装?

在终端执行 npx skills add https://github.com/jarrodmedrano/jarrod-claude-skills --skill game-feel。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

game-feel 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。

game-feel 需要什么权限?

权限级别是执行命令。安装流程涉及命令执行,可能通过 npx skills add https://github.com/jarrodmedrano/jarrod-claude-skills --skill game-feel 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

game-feel 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。

game-feel 用的人多吗?

公开数据是累计安装 499 次,近一周安装 20 次,下载 162 次,GitHub 1 星。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。

game-feel 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

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。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要执行命令,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 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。
  • 需要处理前端设计类任务、又不想每次都手写提示词的人。

不太适合

  • 不能接受 Skill 在你机器上执行命令的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 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
game-feel

游戏感觉 · 本页

4991执行命令未明确通过

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

Game Feel Skill

A framework for understanding, measuring, and creating satisfying game controls based on Steve Swink's comprehensive analysis.

Core Definition

Game feel is the tactile, kinesthetic sense of manipulating a virtual object. It's the sensation of control—that visceral feeling of steering, jumping, and interacting that exists somewhere between player and game.

"Game feel is an invisible art. If a designer's done their job correctly, the player will never notice it. It will just seem right."

The Three Building Blocks

Game feel requires all three elements working together:

ElementDefinitionWithout It
Real-Time ControlContinuous, immediate response to inputFeels like giving orders, not controlling
Simulated SpaceCollision and physics in a virtual worldNo sense of physical interaction
PolishEffects that emphasize interactionsFlat, lifeless, unconvincing

Human Perception Thresholds

The Correction Cycle

Players perceive, think, and act in ~240ms cycles:

  • Perceptual processor: ~100ms
  • Cognitive processor: ~70ms
  • Motor processor: ~70ms

Critical Thresholds

ThresholdValueEffect
Motion illusion10+ fpsBelow this, no sense of movement
Smooth motion30+ fpsMovement feels fluid
Instantaneous response<50msFeels like direct control
Noticeable lag100-200msSluggish but usable
Broken control>240msPlayer notices delay, feel breaks down
"At 50ms response, the game feels like an extension of your body. Above 100ms, you notice lag. Above 240ms, real-time control is broken."

See: references/perception-thresholds.md


The Six Metrics of Game Feel

1. Input

The physical device and signals it sends.

Measure: Sensitivity, states, signal types (Boolean vs continuous), physical ergonomics

2. Response

How input maps to game state changes.

Measure: Direct vs indirect mapping, simulation complexity, ADSR envelopes

3. Context

The spatial environment providing meaning to motion.

Measure: Object spacing relative to avatar speed, collision density, level layout

4. Polish

Effects that emphasize and sell interactions.

Measure: Particles, screen shake, animation sync, sound design

5. Metaphor

What the game represents and expectations it creates.

Measure: Realism vs abstraction, player expectations, genre conventions

6. Rules

Game rules that affect moment-to-moment feel.

Measure: Health systems, risk/reward, state changes, ability unlocks

See: references/six-metrics.md


The ADSR Envelope for Game Feel

Borrowed from audio synthesis, describes how response changes over time:

     Sustain ___________
    /                   \
   / Attack      Decay   \ Release
  /                       \
_/                         \_____
PhaseDescriptionExample (Mario Jump)
AttackTime to reach full responseJump force ramps up as button held
DecaySettling to sustained levelInitial burst settles
SustainMaintained level while input heldMaximum jump height maintained
ReleaseFalloff after input stopsGravity takes over on release

Key insight: Most "floaty" vs "tight" feelings come from Attack and Release times.

See: references/adsr-tuning.md


Common Feel Vocabulary

TermMeaningTypical Cause
TightPrecise, immediate responseShort attack, high acceleration, low release
FloatyLoose, delayed, driftyLong attack/release, high inertia
ResponsiveDoes what you want immediately<100ms response, direct mapping
SluggishDelayed, heavy feeling>150ms response, long attack
SlipperyHard to stop preciselyLow friction, long release
StickyHard to start movingHigh friction, long attack
WeightySense of mass and momentumAcceleration curves, gravity strength
SnappyQuick state transitionsShort attack AND release

Simulation Fundamentals

Position vs Velocity vs Acceleration

LevelWhat ChangesFeel
Set PositionTeleport directlyStiff, robotic (Donkey Kong)
Set VelocityChange speed directlyResponsive but unnatural
Apply ForceAdd accelerationFluid, physical (Mario, Asteroids)

The Asteroids Principle

Separate thrust from rotation for expressive space-feel:

  • Rotation: Direct, instant (no simulation)
  • Thrust: Adds force in facing direction (simulated)
  • Result: Ship feels on-the-edge-of-control but never actually out of control

The Mario Principle

Separate horizontal and vertical systems:

  • Horizontal: Acceleration, max speed, deceleration, run modifier
  • Vertical: Jump force, gravity, fall gravity (3x normal!), terminal velocity
  • Button hold time affects jump height (with min/max limits)

See: references/simulation-recipes.md


The Mario Jump Recipe

The most-analyzed jump in game history:

Horizontal Movement

Walk acceleration → Walk max speed
Run acceleration → Run max speed (B held)
Air acceleration (reduced)
Deceleration (same for walk/run)

Vertical Movement

Initial jump force (instant, large)
Gravity (constant, moderate)
Jump button hold → extends upward force (with max time)
Early release → force artificially set to low value
Apex detection → gravity triples for descent
Terminal velocity cap

The "Hack" That Makes It Work

When player releases jump early:

  1. Check if upward velocity > threshold
  2. If yes, instantly set velocity to preset low value
  3. This creates consistent arc shapes regardless of release timing

See: references/mario-mechanics.md


Polish That Matters

The Three-Tier Impact System

Light / Medium / Hard impacts each get:

  • Distinct animation
  • Distinct sound
  • Distinct visual effect (particles, screen shake)

Sound-Motion Harmony

  • Rising pitch = rising motion (Mario's jump sound)
  • Impact sounds match visual scale
  • Footsteps sync with animation frames
  • Material-specific sounds (metal, grass, stone)

Crossover Sensation

Multiple mechanics should feel bound by the same physics:

  • Swimming feels floaty *because* running feels grounded
  • Flying defies *the same gravity* that affects jumping
  • Contrast creates perceived consistency

See: references/polish-effects.md


Context: Level Design for Feel

Spatial Relationships Must Match Mechanics

  • Jump heights → platform heights
  • Run speed → corridor widths
  • Stopping distance → gap sizes before hazards

The "Just Right" Principle

Objects should be spaced so the intended move *just barely* works:

  • Long Jump gaps: exactly Long Jump distance
  • Triple Jump heights: exactly Triple Jump apex
  • Creates sense of mastery when executed

Soft Boundaries

Use physics to guide, not walls to block:

  • Steep inclines cause sliding (Mario 64)
  • Water slows movement
  • Winds push in intended directions
"Players don't feel the direct intervention of the designer. The limit feels like a logical consequence rather than an overt constraint."

See: references/spatial-context.md


Debugging Feel Problems

SymptomLikely CauseFix
"Floaty"Long attack/release timesShorten acceleration curves
"Sluggish"High response lagReduce input-to-display latency
"Slippery"Low decelerationIncrease friction/deceleration
"Stiff"No acceleration curveAdd attack phase to movement
"Unresponsive"Input not registeringCheck input polling rate
"Unpredictable"Variable trajectoriesUse fixed special-move arcs
"Weightless"Weak gravityIncrease fall gravity especially

Principles of Good Game Feel

  1. Predictable special moves — Fixed trajectories for precision maneuvers
  2. Variable basic moves — Expressive range for moment-to-moment control
  3. Consistent abstraction — Simple physics, but self-consistent
  4. Exceed metaphor expectations — Feel more real than graphics suggest
  5. Polish harmonizes — Sound, visual, animation tell same story
  6. Context matches capability — Level design respects avatar limits
  7. Response < 100ms — Maintain instantaneous feel
  8. Contrast creates variety — Different mechanics feel different

Quick Reference: Classic Feel Profiles

Asteroids (Floaty Space)

  • Thrust separate from rotation
  • Very low friction (4+ seconds to stop)
  • Screen wrap containment
  • Rotation: instant, no simulation

Super Mario Bros (Platformer Gold Standard)

  • Separate horizontal/vertical systems
  • Variable jump height (button hold)
  • Triple gravity on descent
  • Run modifier changes acceleration AND max speed
  • Reduced air control

Mario 64 (3D Translation)

  • Camera-relative thumbstick control
  • Incline-based sliding physics
  • Multiple jump types with fixed trajectories
  • Carving turn interpolation
  • Ground pound as precision landing tool

See: references/classic-profiles.md


Key Mantras

  • "The game should feel like an extension of your body."
  • "Separate systems, consistent world." — Independent mechanics, unified physics.
  • "Exceed the metaphor." — Feel better than it looks.
  • "Fixed for precision, variable for expression." — Special moves vs basic moves.
  • "Polish is not optional." — Effects sell the interaction.
  • "240ms is the wall." — Response time ceiling for real-time control.

平台分布

Codex

35.07%
按下载量换算57

Claude

28.19%
按下载量换算46

Cursor

19.14%
按下载量换算31

Gemini CLI

10.16%
按下载量换算16

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

执行命令

安装流程涉及命令执行,可能通过 npx skills add https://github.com/jarrodmedrano/jarrod-claude-skills --skill game-feel 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

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