Token导航 LogoToken导航

Agent Skill · 前端设计

desk-accessory(办公桌配件)

desk-accessory 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。

desk-accessory 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,权限为调用外部服务;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 7,463。

总安装 372周安装 16GitHub Stars 309下载 131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:desk-accessory(办公桌配件)
来源仓库:https://github.com/paulrobello/claude-office
仓库路径:skills/desk-accessory
安装命令:
npx skills add https://github.com/paulrobello/claude-office --skill desk-accessory
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/paulrobello/claude-office --skill desk-accessory

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。

该技能生成办公桌配件素材,基于 Nano Banana MCP 与 ImageMagick 处理。提倡灰度或白色设计以便运行时染色,增加视觉多样性而不重复生成。区分独特物品与常见物品,前者需彩色设计,后者保持单色以节省资源。安装前需确认图像生成服务可用性,注意输出为透明 PNG,建议本地调色测试。

desk-accessory 属于前端设计类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/desk-accessory

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/paulrobello/claude-office --skill desk-accesso…;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

    它需要调用外部服务,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

desk-accessory 怎么安装?

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

desk-accessory 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。

desk-accessory 需要什么权限?

权限级别是调用外部服务。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。

desk-accessory 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。

desk-accessory 用的人多吗?

公开数据是累计安装 372 次,近一周安装 16 次,下载 131 次,GitHub 309 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。

desk-accessory 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

  • 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

  • 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
desk-accessory

办公桌配件 · 本页

372309调用外部服务未明确通过

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

Desk Accessory Generator

Create desk accessories for the Claude Office Visualizer using Nano Banana MCP for AI image generation and ImageMagick for transparency and tinting preparation.

Philosophy: Tintable Variety

Desk accessories should be grayscale or white so they can be tinted with different colors at runtime. This allows a single sprite to represent many color variations across different desks, adding visual variety without generating multiple sprites.

Before generating, ask:

  • Is this a unique item (one-of-a-kind) or a common item (appears on multiple desks)?
  • If common: make it WHITE/GRAYSCALE for tinting
  • If unique: colors can be baked in

Key Principle: White for Tinting

CRITICAL: Unless the item is unique (only appears once), generate it as WHITE or GRAYSCALE so it can be tinted programmatically.

Why this matters:

  • A white coffee mug can become blue, pink, green, gold via tinting
  • A red stapler can ONLY be red - no variety
  • Tinting multiplies the sprite's colors with the tint color
  • White (0xFFFFFF) × any tint = that tint color
  • Gray gives shaded/dimensional tinting results

Key Principle: Facing Direction

CRITICAL: All desk items should face NORTH-WEST (away from the viewer). The game uses a top-down perspective where the user views characters and items from behind/above. Items should be oriented so their "front" faces away from the camera toward the top-left of the screen.

Prompt Templates

Scaling Guidance

IMPORTANT: Generated images (1024px) are scaled down to ~128px for use in the game. Include guidance in your prompts to ensure details survive scaling:

  • Request thick lines and bold shapes
  • Ask for chunky pixel art style
  • Avoid fine details that will disappear at small sizes
  • Add explicit note: "this will be scaled down significantly, use thick prominent features"

Common Desk Items (Tintable - WHITE/GRAYSCALE)

Pixel art [ITEM], 16-bit retro game style, WHITE colored, simple clean design,
facing north-west, on solid magenta background #FF00FF, game sprite asset,
centered composition, no shadows on background, clean edges,
chunky pixel art with thick lines and bold shapes, this will be scaled down significantly so use thick prominent features

Tintable item examples:

  • Coffee mug (white ceramic)
  • Stapler (white/gray body)
  • Tape dispenser (white/gray)
  • Pencil cup (white ceramic)
  • Small plant pot (white ceramic)
  • Desk lamp (white/gray)
  • Paper tray (white/gray)

Unique Desk Items (Full Color OK)

Pixel art [ITEM], 16-bit retro game style, [SPECIFIC COLORS], simple clean design,
facing north-west, on solid magenta background #FF00FF, game sprite asset,
centered composition, no shadows on background, clean edges,
chunky pixel art with thick lines and bold shapes, this will be scaled down significantly so use thick prominent features

Unique item examples (color OK):

  • Globe (blue/green earth colors)
  • Family photo frame (wood frame, photo inside)
  • Award trophy (gold)
  • Specific branded items

Processing Workflow

Step 1: Generate with Nano Banana

Use mcpl to call the nanobanana generate_image tool. Always use model_tier: "pro" and resolution: "1k":

mcpl call nanobanana generate_image '{"prompt": "Pixel art white coffee mug, 16-bit retro game style, WHITE colored ceramic, simple clean design, facing north-west, on solid magenta background #FF00FF, game sprite asset, centered composition, no shadows on background, chunky pixel art with thick lines and bold shapes, this will be scaled down significantly so use thick prominent features", "model_tier": "pro", "resolution": "1k"}'

Why these settings:

  • model_tier: "pro": Best quality generation
  • resolution: "1k": Smallest output (1024px) since we're scaling down anyway - saves processing time
  • Prompt includes scaling guidance so the AI uses thick, prominent features that survive downscaling

Step 2: Process for Transparency + Grayscale + Scaling

For tintable items, use the helper script:

# Default: scales to 128px max dimension (ideal for desk accessories)
.claude/skills/desk-accessory/scripts/process_tintable.sh input.png output.png

# Custom size for larger/more detailed items
.claude/skills/desk-accessory/scripts/process_tintable.sh input.png output.png 15 140 192

# Skip scaling (keep original ~1024px size)
.claude/skills/desk-accessory/scripts/process_tintable.sh input.png output.png 15 140 0

Arguments: input.png output.png [fuzz%] [brightness%] [target_size]

  • fuzz%: Background color tolerance (default: 15)
  • brightness%: Brightness adjustment (default: 140)
  • target_size: Max dimension in pixels (default: 128, use 0 to skip)

The script now uses a multi-pass workflow for thorough magenta removal:

  1. FFmpeg geq filter: Removes pixels where R≈B and G is low (purple/magenta hues)
  2. ImageMagick multi-pass: Catches remaining bright magenta shades (#FF00FF, #CC00CC, etc.)
  3. ImageMagick dark purple cleanup: Removes dark edge pixels like rgb(32,0,31)
  4. Desaturation: Converts to grayscale for tinting
  5. Scaling: Uses -filter Point for pixel-art-friendly nearest-neighbor scaling

Manual workflow (if scripts unavailable):

INPUT="generated.png"
OUTPUT="accessory.png"
TARGET_SIZE=128

# Step 1: FFmpeg - remove purple hue pixels
ffmpeg -y -i "$INPUT" \
  -vf "geq=r='r(X,Y)':g='g(X,Y)':b='b(X,Y)':a='if(between(r(X,Y)-b(X,Y),-60,60)*lt(g(X,Y),r(X,Y)*0.7)*lt(g(X,Y),b(X,Y)*0.7)*gt(r(X,Y)+b(X,Y),100),0,alpha(X,Y))'" \
  -update 1 -frames:v 1 /tmp/step1.png

# Step 2: ImageMagick - remove remaining magenta shades
magick /tmp/step1.png -alpha set -channel RGBA \
  -fuzz 20% -transparent "magenta" \
  -fuzz 15% -transparent "#CC00CC" \
  -fuzz 15% -transparent "#880088" \
  /tmp/step2.png

# Step 3: Dark purple cleanup + desaturate + scale
magick /tmp/step2.png \
  -fuzz 8% -fill none -opaque "rgb(32,0,31)" -opaque "rgb(34,0,31)" \
  -modulate 140,0,100 \
  -trim +repage \
  -filter Point -resize "${TARGET_SIZE}x${TARGET_SIZE}>" \
  -type TrueColorAlpha -strip \
  "$OUTPUT"

Key flags explained:

  • -modulate 140,0,100: Brightness 140%, Saturation 0% (grayscale), Hue unchanged
  • -filter Point: Nearest-neighbor interpolation (preserves pixel art crispness)
  • -resize "128x128>": Scale down to max 128px, only if larger (the > suffix)
  • -type TrueColorAlpha: Ensure RGBA format for PixiJS compatibility

Why multi-pass? AI-generated images often have anti-aliased edges that blend the subject with the magenta background. Single-pass removal leaves pink/purple fringing. The multi-pass approach targets different shades of magenta/purple for thorough cleanup

Step 3: Verify Result

# Check has transparency and correct format
magick "$OUTPUT" -format "Size: %wx%h, Channels: %[channels], Transparent: %[opaque]" info:
# Expected: Channels: srgba 4.0, Transparent: False

Step 4: Place in Project

mv "$OUTPUT" frontend/public/sprites/

Integration with OfficeGameV2

Desk accessories are rendered in DeskSurfaces component with tinting:

// Define tint colors array
const accessoryTints = [
  0xffffff, // White (no tint)
  0x87ceeb, // Sky blue
  0x98fb98, // Pale green
  0xffb6c1, // Light pink
  0xffd700, // Gold
  0xdda0dd, // Plum
  0xf0e68c, // Khaki
  0xadd8e6, // Light blue
];

// Apply tint based on desk index
<pixiSprite
  texture={accessoryTexture}
  tint={accessoryTints[deskIndex % accessoryTints.length]}
  ...
/>

Adding New Accessories

  1. Add texture state in OfficeGameV2:
const [newItemTexture, setNewItemTexture] = useState<Texture | null>(null);
  1. Load in useEffect:
Assets.load("/sprites/new-item.png"),
  1. Add to DeskSurfacesProps interface and function parameters
  2. Render with tint:
{newItemTexture && (
  <pixiSprite
    texture={newItemTexture}
    anchor={0.5}
    x={POSITION_X}
    y={POSITION_Y}
    scale={SCALE}
    tint={accessoryTints[i % accessoryTints.length]}
  />
)}

Randomization Pattern

To randomize which desks get which items:

// Deterministic "random" based on desk index
const getDeskItem = (index: number): "mug" | "stapler" | "plant" => {
  const items = ["mug", "stapler", "plant"] as const;
  return items[index % items.length];
};

Or use a seeded random for more variety:

// Simple hash for deterministic randomness
const getDeskItem = (index: number): string => {
  const items = ["mug", "stapler", "plant", "lamp"];
  const hash = (index * 2654435761) % items.length;
  return items[Math.abs(hash)];
};

Output Location

Default: frontend/public/sprites/

Naming convention: lowercase-with-dashes.png

  • coffee-mug.png
  • stapler.png
  • desk-plant.png
  • tape-dispenser.png

Quick Reference

# 1. Generate white/grayscale item - use pro model at 1k resolution
mcpl call nanobanana generate_image '{"prompt": "Pixel art white [ITEM], 16-bit retro game style, WHITE colored, simple clean design, facing north-west, on solid magenta background #FF00FF, game sprite asset, centered composition, clean edges, chunky pixel art with thick lines and bold shapes, this will be scaled down significantly so use thick prominent features", "model_tier": "pro", "resolution": "1k"}'

# 2. Process for tinting (default: scales to 128px, grayscale, transparent background)
.claude/skills/desk-accessory/scripts/process_tintable.sh /path/to/generated.png frontend/public/sprites/item-name.png

# 2b. For larger/more detailed items, use a larger target size
.claude/skills/desk-accessory/scripts/process_tintable.sh /path/to/generated.png frontend/public/sprites/item-name.png 15 140 192

# 3. Verify
magick frontend/public/sprites/item-name.png -format "%wx%h %[channels] %[opaque]" info:
# Should show: ~128x128 (or smaller) srgba 4.0 False

Anti-Patterns

❌ Generating colored common items Why: Can't be tinted to other colors effectively. Better: Generate white/grayscale, tint at runtime.

❌ Using flood fill for items with holes Why: Flood fill from corners won't reach enclosed areas (like mug handles). Better: Use global -transparent for items with enclosed spaces.

❌ Forgetting -type TrueColorAlpha Why: Grayscale conversion creates 2-channel images that may not render correctly. Better: Always convert back to RGBA after desaturation.

❌ Making items too detailed Why: Small desk items are rendered at tiny scales (~0.025-0.03). Better: Simple, bold shapes that read well at small sizes.

Existing Accessories

ItemFileTintableNotes
Coffee Mugcoffee-mug.pngYesWhite ceramic, no steam
Staplerstapler.pngYesGrayscale, brightened
Desk Lampdesk-lamp.pngYesAdjustable lamp, facing NW

Remember

  • Common items = WHITE/GRAYSCALE for tinting variety
  • Process with -modulate X,0,100 to desaturate
  • Always use -type TrueColorAlpha for PixiJS compatibility
  • Test tinting looks good with multiple colors before committing

平台分布

Codex

36.61%
按下载量换算48

Claude

29.42%
按下载量换算39

Cursor

20.12%
按下载量换算26

Gemini CLI

8.58%
按下载量换算11

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

调用外部服务

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。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